admin/scribe
1
0
Fork 0
Code Issues Pull Requests Actions 3 Packages Projects Releases Wiki Activity

Merge pull request #14 from JSchmie/develop

Browse Source 
Develop
This commit is contained in:
Jacob Schmieder
2023-09-22 19:26:29 +02:00
committed by GitHub
parent 950f0834ef 0be90cd9ee
commit 296efecd67
40 changed files with 3674 additions and 755 deletions
Download Patch File Download Diff File Expand all files Collapse all files
.github/workflows/python-publish.yml
+39
View File
@@ -0,0 +1,39 @@
# This workflow will upload a Python Package using Twine when a release is created
# For more information see: https://docs.github.com/en/actions/automating-builds-and-tests/building-and-testing-python#publishing-to-package-registries
# This workflow uses actions that are not certified by GitHub.
# They are provided by a third-party and are governed by
# separate terms of service, privacy policy, and support
# documentation.
name: Upload Python Package
on:
release:
types: [published]
permissions:
contents: read
jobs:
deploy:
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v3
- name: Set up Python
uses: actions/setup-python@v3
with:
python-version: '3.x'
- name: Install dependencies
run: |
python -m pip install --upgrade pip
pip install build
- name: Build package
run: python -m build
- name: Publish package
uses: pypa/gh-action-pypi-publish@27b31702a0e7fc50959f5ad993c78deac1bdfc29
with:
user: __token__
password: ${{ secrets.PYPI_API_TOKEN }}
Dockerfile
+26
View File
@@ -0,0 +1,26 @@
#pytorch Image
FROM pytorch/pytorch:1.11.0-cuda11.3-cudnn8-runtime
# Install dependencies
WORKDIR /app
ARG hf_token
#Enviorment Dependncies
ENV TRANSFORMERS_CACHE /app/models
ENV HF_HOME /app/models
ENV AUTOT_CACHE /app/models
ENV PYANNOTE_CACHE /app/models/pyannote
#Copy all necessary files
COPY requirements.txt /app/requirements.txt
COPY scraibe /app/Scraibe
COPY setup.py /app/setup.py
#Installing all necessary Dependencies and Running the Application with a personalised Hugging-Face-Token
RUN conda install pip
RUN conda install -y ffmpeg
RUN conda install -c conda-forge libsndfile
RUN pip install torchaudio==0.11.0+cu113 -f https://download.pytorch.org/whl/torch_stable.html
RUN pip install /app/
RUN pip install markupsafe==2.0.1 --force-reinstall
RUN Scraibe --hf_token $hf_token
# Expose port
EXPOSE 7860
# Run the application
ENTRYPOINT ["scraibe"]
LICENSE
+674
View File
@@ -0,0 +1,674 @@
GNU GENERAL PUBLIC LICENSE
Version 3, 29 June 2007
Copyright (C) 2007 Free Software Foundation, Inc. <https://fsf.org/>
Everyone is permitted to copy and distribute verbatim copies
of this license document, but changing it is not allowed.
Preamble
The GNU General Public License is a free, copyleft license for
software and other kinds of works.
The licenses for most software and other practical works are designed
to take away your freedom to share and change the works. By contrast,
the GNU General Public License is intended to guarantee your freedom to
share and change all versions of a program--to make sure it remains free
software for all its users. We, the Free Software Foundation, use the
GNU General Public License for most of our software; it applies also to
any other work released this way by its authors. You can apply it to
your programs, too.
When we speak of free software, we are referring to freedom, not
price. Our General Public Licenses are designed to make sure that you
have the freedom to distribute copies of free software (and charge for
them if you wish), that you receive source code or can get it if you
want it, that you can change the software or use pieces of it in new
free programs, and that you know you can do these things.
To protect your rights, we need to prevent others from denying you
these rights or asking you to surrender the rights. Therefore, you have
certain responsibilities if you distribute copies of the software, or if
you modify it: responsibilities to respect the freedom of others.
For example, if you distribute copies of such a program, whether
gratis or for a fee, you must pass on to the recipients the same
freedoms that you received. You must make sure that they, too, receive
or can get the source code. And you must show them these terms so they
know their rights.
Developers that use the GNU GPL protect your rights with two steps:
(1) assert copyright on the software, and (2) offer you this License
giving you legal permission to copy, distribute and/or modify it.
For the developers' and authors' protection, the GPL clearly explains
that there is no warranty for this free software. For both users' and
authors' sake, the GPL requires that modified versions be marked as
changed, so that their problems will not be attributed erroneously to
authors of previous versions.
Some devices are designed to deny users access to install or run
modified versions of the software inside them, although the manufacturer
can do so. This is fundamentally incompatible with the aim of
protecting users' freedom to change the software. The systematic
pattern of such abuse occurs in the area of products for individuals to
use, which is precisely where it is most unacceptable. Therefore, we
have designed this version of the GPL to prohibit the practice for those
products. If such problems arise substantially in other domains, we
stand ready to extend this provision to those domains in future versions
of the GPL, as needed to protect the freedom of users.
Finally, every program is threatened constantly by software patents.
States should not allow patents to restrict development and use of
software on general-purpose computers, but in those that do, we wish to
avoid the special danger that patents applied to a free program could
make it effectively proprietary. To prevent this, the GPL assures that
patents cannot be used to render the program non-free.
The precise terms and conditions for copying, distribution and
modification follow.
TERMS AND CONDITIONS
0. Definitions.
"This License" refers to version 3 of the GNU General Public License.
"Copyright" also means copyright-like laws that apply to other kinds of
works, such as semiconductor masks.
"The Program" refers to any copyrightable work licensed under this
License. Each licensee is addressed as "you". "Licensees" and
"recipients" may be individuals or organizations.
To "modify" a work means to copy from or adapt all or part of the work
in a fashion requiring copyright permission, other than the making of an
exact copy. The resulting work is called a "modified version" of the
earlier work or a work "based on" the earlier work.
A "covered work" means either the unmodified Program or a work based
on the Program.
To "propagate" a work means to do anything with it that, without
permission, would make you directly or secondarily liable for
infringement under applicable copyright law, except executing it on a
computer or modifying a private copy. Propagation includes copying,
distribution (with or without modification), making available to the
public, and in some countries other activities as well.
To "convey" a work means any kind of propagation that enables other
parties to make or receive copies. Mere interaction with a user through
a computer network, with no transfer of a copy, is not conveying.
An interactive user interface displays "Appropriate Legal Notices"
to the extent that it includes a convenient and prominently visible
feature that (1) displays an appropriate copyright notice, and (2)
tells the user that there is no warranty for the work (except to the
extent that warranties are provided), that licensees may convey the
work under this License, and how to view a copy of this License. If
the interface presents a list of user commands or options, such as a
menu, a prominent item in the list meets this criterion.
1. Source Code.
The "source code" for a work means the preferred form of the work
for making modifications to it. "Object code" means any non-source
form of a work.
A "Standard Interface" means an interface that either is an official
standard defined by a recognized standards body, or, in the case of
interfaces specified for a particular programming language, one that
is widely used among developers working in that language.
The "System Libraries" of an executable work include anything, other
than the work as a whole, that (a) is included in the normal form of
packaging a Major Component, but which is not part of that Major
Component, and (b) serves only to enable use of the work with that
Major Component, or to implement a Standard Interface for which an
implementation is available to the public in source code form. A
"Major Component", in this context, means a major essential component
(kernel, window system, and so on) of the specific operating system
(if any) on which the executable work runs, or a compiler used to
produce the work, or an object code interpreter used to run it.
The "Corresponding Source" for a work in object code form means all
the source code needed to generate, install, and (for an executable
work) run the object code and to modify the work, including scripts to
control those activities. However, it does not include the work's
System Libraries, or general-purpose tools or generally available free
programs which are used unmodified in performing those activities but
which are not part of the work. For example, Corresponding Source
includes interface definition files associated with source files for
the work, and the source code for shared libraries and dynamically
linked subprograms that the work is specifically designed to require,
such as by intimate data communication or control flow between those
subprograms and other parts of the work.
The Corresponding Source need not include anything that users
can regenerate automatically from other parts of the Corresponding
Source.
The Corresponding Source for a work in source code form is that
same work.
2. Basic Permissions.
All rights granted under this License are granted for the term of
copyright on the Program, and are irrevocable provided the stated
conditions are met. This License explicitly affirms your unlimited
permission to run the unmodified Program. The output from running a
covered work is covered by this License only if the output, given its
content, constitutes a covered work. This License acknowledges your
rights of fair use or other equivalent, as provided by copyright law.
You may make, run and propagate covered works that you do not
convey, without conditions so long as your license otherwise remains
in force. You may convey covered works to others for the sole purpose
of having them make modifications exclusively for you, or provide you
with facilities for running those works, provided that you comply with
the terms of this License in conveying all material for which you do
not control copyright. Those thus making or running the covered works
for you must do so exclusively on your behalf, under your direction
and control, on terms that prohibit them from making any copies of
your copyrighted material outside their relationship with you.
Conveying under any other circumstances is permitted solely under
the conditions stated below. Sublicensing is not allowed; section 10
makes it unnecessary.
3. Protecting Users' Legal Rights From Anti-Circumvention Law.
No covered work shall be deemed part of an effective technological
measure under any applicable law fulfilling obligations under article
11 of the WIPO copyright treaty adopted on 20 December 1996, or
similar laws prohibiting or restricting circumvention of such
measures.
When you convey a covered work, you waive any legal power to forbid
circumvention of technological measures to the extent such circumvention
is effected by exercising rights under this License with respect to
the covered work, and you disclaim any intention to limit operation or
modification of the work as a means of enforcing, against the work's
users, your or third parties' legal rights to forbid circumvention of
technological measures.
4. Conveying Verbatim Copies.
You may convey verbatim copies of the Program's source code as you
receive it, in any medium, provided that you conspicuously and
appropriately publish on each copy an appropriate copyright notice;
keep intact all notices stating that this License and any
non-permissive terms added in accord with section 7 apply to the code;
keep intact all notices of the absence of any warranty; and give all
recipients a copy of this License along with the Program.
You may charge any price or no price for each copy that you convey,
and you may offer support or warranty protection for a fee.
5. Conveying Modified Source Versions.
You may convey a work based on the Program, or the modifications to
produce it from the Program, in the form of source code under the
terms of section 4, provided that you also meet all of these conditions:
a) The work must carry prominent notices stating that you modified
it, and giving a relevant date.
b) The work must carry prominent notices stating that it is
released under this License and any conditions added under section
7. This requirement modifies the requirement in section 4 to
"keep intact all notices".
c) You must license the entire work, as a whole, under this
License to anyone who comes into possession of a copy. This
License will therefore apply, along with any applicable section 7
additional terms, to the whole of the work, and all its parts,
regardless of how they are packaged. This License gives no
permission to license the work in any other way, but it does not
invalidate such permission if you have separately received it.
d) If the work has interactive user interfaces, each must display
Appropriate Legal Notices; however, if the Program has interactive
interfaces that do not display Appropriate Legal Notices, your
work need not make them do so.
A compilation of a covered work with other separate and independent
works, which are not by their nature extensions of the covered work,
and which are not combined with it such as to form a larger program,
in or on a volume of a storage or distribution medium, is called an
"aggregate" if the compilation and its resulting copyright are not
used to limit the access or legal rights of the compilation's users
beyond what the individual works permit. Inclusion of a covered work
in an aggregate does not cause this License to apply to the other
parts of the aggregate.
6. Conveying Non-Source Forms.
You may convey a covered work in object code form under the terms
of sections 4 and 5, provided that you also convey the
machine-readable Corresponding Source under the terms of this License,
in one of these ways:
a) Convey the object code in, or embodied in, a physical product
(including a physical distribution medium), accompanied by the
Corresponding Source fixed on a durable physical medium
customarily used for software interchange.
b) Convey the object code in, or embodied in, a physical product
(including a physical distribution medium), accompanied by a
written offer, valid for at least three years and valid for as
long as you offer spare parts or customer support for that product
model, to give anyone who possesses the object code either (1) a
copy of the Corresponding Source for all the software in the
product that is covered by this License, on a durable physical
medium customarily used for software interchange, for a price no
more than your reasonable cost of physically performing this
conveying of source, or (2) access to copy the
Corresponding Source from a network server at no charge.
c) Convey individual copies of the object code with a copy of the
written offer to provide the Corresponding Source. This
alternative is allowed only occasionally and noncommercially, and
only if you received the object code with such an offer, in accord
with subsection 6b.
d) Convey the object code by offering access from a designated
place (gratis or for a charge), and offer equivalent access to the
Corresponding Source in the same way through the same place at no
further charge. You need not require recipients to copy the
Corresponding Source along with the object code. If the place to
copy the object code is a network server, the Corresponding Source
may be on a different server (operated by you or a third party)
that supports equivalent copying facilities, provided you maintain
clear directions next to the object code saying where to find the
Corresponding Source. Regardless of what server hosts the
Corresponding Source, you remain obligated to ensure that it is
available for as long as needed to satisfy these requirements.
e) Convey the object code using peer-to-peer transmission, provided
you inform other peers where the object code and Corresponding
Source of the work are being offered to the general public at no
charge under subsection 6d.
A separable portion of the object code, whose source code is excluded
from the Corresponding Source as a System Library, need not be
included in conveying the object code work.
A "User Product" is either (1) a "consumer product", which means any
tangible personal property which is normally used for personal, family,
or household purposes, or (2) anything designed or sold for incorporation
into a dwelling. In determining whether a product is a consumer product,
doubtful cases shall be resolved in favor of coverage. For a particular
product received by a particular user, "normally used" refers to a
typical or common use of that class of product, regardless of the status
of the particular user or of the way in which the particular user
actually uses, or expects or is expected to use, the product. A product
is a consumer product regardless of whether the product has substantial
commercial, industrial or non-consumer uses, unless such uses represent
the only significant mode of use of the product.
"Installation Information" for a User Product means any methods,
procedures, authorization keys, or other information required to install
and execute modified versions of a covered work in that User Product from
a modified version of its Corresponding Source. The information must
suffice to ensure that the continued functioning of the modified object
code is in no case prevented or interfered with solely because
modification has been made.
If you convey an object code work under this section in, or with, or
specifically for use in, a User Product, and the conveying occurs as
part of a transaction in which the right of possession and use of the
User Product is transferred to the recipient in perpetuity or for a
fixed term (regardless of how the transaction is characterized), the
Corresponding Source conveyed under this section must be accompanied
by the Installation Information. But this requirement does not apply
if neither you nor any third party retains the ability to install
modified object code on the User Product (for example, the work has
been installed in ROM).
The requirement to provide Installation Information does not include a
requirement to continue to provide support service, warranty, or updates
for a work that has been modified or installed by the recipient, or for
the User Product in which it has been modified or installed. Access to a
network may be denied when the modification itself materially and
adversely affects the operation of the network or violates the rules and
protocols for communication across the network.
Corresponding Source conveyed, and Installation Information provided,
in accord with this section must be in a format that is publicly
documented (and with an implementation available to the public in
source code form), and must require no special password or key for
unpacking, reading or copying.
7. Additional Terms.
"Additional permissions" are terms that supplement the terms of this
License by making exceptions from one or more of its conditions.
Additional permissions that are applicable to the entire Program shall
be treated as though they were included in this License, to the extent
that they are valid under applicable law. If additional permissions
apply only to part of the Program, that part may be used separately
under those permissions, but the entire Program remains governed by
this License without regard to the additional permissions.
When you convey a copy of a covered work, you may at your option
remove any additional permissions from that copy, or from any part of
it. (Additional permissions may be written to require their own
removal in certain cases when you modify the work.) You may place
additional permissions on material, added by you to a covered work,
for which you have or can give appropriate copyright permission.
Notwithstanding any other provision of this License, for material you
add to a covered work, you may (if authorized by the copyright holders of
that material) supplement the terms of this License with terms:
a) Disclaiming warranty or limiting liability differently from the
terms of sections 15 and 16 of this License; or
b) Requiring preservation of specified reasonable legal notices or
author attributions in that material or in the Appropriate Legal
Notices displayed by works containing it; or
c) Prohibiting misrepresentation of the origin of that material, or
requiring that modified versions of such material be marked in
reasonable ways as different from the original version; or
d) Limiting the use for publicity purposes of names of licensors or
authors of the material; or
e) Declining to grant rights under trademark law for use of some
trade names, trademarks, or service marks; or
f) Requiring indemnification of licensors and authors of that
material by anyone who conveys the material (or modified versions of
it) with contractual assumptions of liability to the recipient, for
any liability that these contractual assumptions directly impose on
those licensors and authors.
All other non-permissive additional terms are considered "further
restrictions" within the meaning of section 10. If the Program as you
received it, or any part of it, contains a notice stating that it is
governed by this License along with a term that is a further
restriction, you may remove that term. If a license document contains
a further restriction but permits relicensing or conveying under this
License, you may add to a covered work material governed by the terms
of that license document, provided that the further restriction does
not survive such relicensing or conveying.
If you add terms to a covered work in accord with this section, you
must place, in the relevant source files, a statement of the
additional terms that apply to those files, or a notice indicating
where to find the applicable terms.
Additional terms, permissive or non-permissive, may be stated in the
form of a separately written license, or stated as exceptions;
the above requirements apply either way.
8. Termination.
You may not propagate or modify a covered work except as expressly
provided under this License. Any attempt otherwise to propagate or
modify it is void, and will automatically terminate your rights under
this License (including any patent licenses granted under the third
paragraph of section 11).
However, if you cease all violation of this License, then your
license from a particular copyright holder is reinstated (a)
provisionally, unless and until the copyright holder explicitly and
finally terminates your license, and (b) permanently, if the copyright
holder fails to notify you of the violation by some reasonable means
prior to 60 days after the cessation.
Moreover, your license from a particular copyright holder is
reinstated permanently if the copyright holder notifies you of the
violation by some reasonable means, this is the first time you have
received notice of violation of this License (for any work) from that
copyright holder, and you cure the violation prior to 30 days after
your receipt of the notice.
Termination of your rights under this section does not terminate the
licenses of parties who have received copies or rights from you under
this License. If your rights have been terminated and not permanently
reinstated, you do not qualify to receive new licenses for the same
material under section 10.
9. Acceptance Not Required for Having Copies.
You are not required to accept this License in order to receive or
run a copy of the Program. Ancillary propagation of a covered work
occurring solely as a consequence of using peer-to-peer transmission
to receive a copy likewise does not require acceptance. However,
nothing other than this License grants you permission to propagate or
modify any covered work. These actions infringe copyright if you do
not accept this License. Therefore, by modifying or propagating a
covered work, you indicate your acceptance of this License to do so.
10. Automatic Licensing of Downstream Recipients.
Each time you convey a covered work, the recipient automatically
receives a license from the original licensors, to run, modify and
propagate that work, subject to this License. You are not responsible
for enforcing compliance by third parties with this License.
An "entity transaction" is a transaction transferring control of an
organization, or substantially all assets of one, or subdividing an
organization, or merging organizations. If propagation of a covered
work results from an entity transaction, each party to that
transaction who receives a copy of the work also receives whatever
licenses to the work the party's predecessor in interest had or could
give under the previous paragraph, plus a right to possession of the
Corresponding Source of the work from the predecessor in interest, if
the predecessor has it or can get it with reasonable efforts.
You may not impose any further restrictions on the exercise of the
rights granted or affirmed under this License. For example, you may
not impose a license fee, royalty, or other charge for exercise of
rights granted under this License, and you may not initiate litigation
(including a cross-claim or counterclaim in a lawsuit) alleging that
any patent claim is infringed by making, using, selling, offering for
sale, or importing the Program or any portion of it.
11. Patents.
A "contributor" is a copyright holder who authorizes use under this
License of the Program or a work on which the Program is based. The
work thus licensed is called the contributor's "contributor version".
A contributor's "essential patent claims" are all patent claims
owned or controlled by the contributor, whether already acquired or
hereafter acquired, that would be infringed by some manner, permitted
by this License, of making, using, or selling its contributor version,
but do not include claims that would be infringed only as a
consequence of further modification of the contributor version. For
purposes of this definition, "control" includes the right to grant
patent sublicenses in a manner consistent with the requirements of
this License.
Each contributor grants you a non-exclusive, worldwide, royalty-free
patent license under the contributor's essential patent claims, to
make, use, sell, offer for sale, import and otherwise run, modify and
propagate the contents of its contributor version.
In the following three paragraphs, a "patent license" is any express
agreement or commitment, however denominated, not to enforce a patent
(such as an express permission to practice a patent or covenant not to
sue for patent infringement). To "grant" such a patent license to a
party means to make such an agreement or commitment not to enforce a
patent against the party.
If you convey a covered work, knowingly relying on a patent license,
and the Corresponding Source of the work is not available for anyone
to copy, free of charge and under the terms of this License, through a
publicly available network server or other readily accessible means,
then you must either (1) cause the Corresponding Source to be so
available, or (2) arrange to deprive yourself of the benefit of the
patent license for this particular work, or (3) arrange, in a manner
consistent with the requirements of this License, to extend the patent
license to downstream recipients. "Knowingly relying" means you have
actual knowledge that, but for the patent license, your conveying the
covered work in a country, or your recipient's use of the covered work
in a country, would infringe one or more identifiable patents in that
country that you have reason to believe are valid.
If, pursuant to or in connection with a single transaction or
arrangement, you convey, or propagate by procuring conveyance of, a
covered work, and grant a patent license to some of the parties
receiving the covered work authorizing them to use, propagate, modify
or convey a specific copy of the covered work, then the patent license
you grant is automatically extended to all recipients of the covered
work and works based on it.
A patent license is "discriminatory" if it does not include within
the scope of its coverage, prohibits the exercise of, or is
conditioned on the non-exercise of one or more of the rights that are
specifically granted under this License. You may not convey a covered
work if you are a party to an arrangement with a third party that is
in the business of distributing software, under which you make payment
to the third party based on the extent of your activity of conveying
the work, and under which the third party grants, to any of the
parties who would receive the covered work from you, a discriminatory
patent license (a) in connection with copies of the covered work
conveyed by you (or copies made from those copies), or (b) primarily
for and in connection with specific products or compilations that
contain the covered work, unless you entered into that arrangement,
or that patent license was granted, prior to 28 March 2007.
Nothing in this License shall be construed as excluding or limiting
any implied license or other defenses to infringement that may
otherwise be available to you under applicable patent law.
12. No Surrender of Others' Freedom.
If conditions are imposed on you (whether by court order, agreement or
otherwise) that contradict the conditions of this License, they do not
excuse you from the conditions of this License. If you cannot convey a
covered work so as to satisfy simultaneously your obligations under this
License and any other pertinent obligations, then as a consequence you may
not convey it at all. For example, if you agree to terms that obligate you
to collect a royalty for further conveying from those to whom you convey
the Program, the only way you could satisfy both those terms and this
License would be to refrain entirely from conveying the Program.
13. Use with the GNU Affero General Public License.
Notwithstanding any other provision of this License, you have
permission to link or combine any covered work with a work licensed
under version 3 of the GNU Affero General Public License into a single
combined work, and to convey the resulting work. The terms of this
License will continue to apply to the part which is the covered work,
but the special requirements of the GNU Affero General Public License,
section 13, concerning interaction through a network will apply to the
combination as such.
14. Revised Versions of this License.
The Free Software Foundation may publish revised and/or new versions of
the GNU General Public License from time to time. Such new versions will
be similar in spirit to the present version, but may differ in detail to
address new problems or concerns.
Each version is given a distinguishing version number. If the
Program specifies that a certain numbered version of the GNU General
Public License "or any later version" applies to it, you have the
option of following the terms and conditions either of that numbered
version or of any later version published by the Free Software
Foundation. If the Program does not specify a version number of the
GNU General Public License, you may choose any version ever published
by the Free Software Foundation.
If the Program specifies that a proxy can decide which future
versions of the GNU General Public License can be used, that proxy's
public statement of acceptance of a version permanently authorizes you
to choose that version for the Program.
Later license versions may give you additional or different
permissions. However, no additional obligations are imposed on any
author or copyright holder as a result of your choosing to follow a
later version.
15. Disclaimer of Warranty.
THERE IS NO WARRANTY FOR THE PROGRAM, TO THE EXTENT PERMITTED BY
APPLICABLE LAW. EXCEPT WHEN OTHERWISE STATED IN WRITING THE COPYRIGHT
HOLDERS AND/OR OTHER PARTIES PROVIDE THE PROGRAM "AS IS" WITHOUT WARRANTY
OF ANY KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING, BUT NOT LIMITED TO,
THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
PURPOSE. THE ENTIRE RISK AS TO THE QUALITY AND PERFORMANCE OF THE PROGRAM
IS WITH YOU. SHOULD THE PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF
ALL NECESSARY SERVICING, REPAIR OR CORRECTION.
16. Limitation of Liability.
IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING
WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MODIFIES AND/OR CONVEYS
THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES, INCLUDING ANY
GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING OUT OF THE
USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED TO LOSS OF
DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY YOU OR THIRD
PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER PROGRAMS),
EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE POSSIBILITY OF
SUCH DAMAGES.
17. Interpretation of Sections 15 and 16.
If the disclaimer of warranty and limitation of liability provided
above cannot be given local legal effect according to their terms,
reviewing courts shall apply local law that most closely approximates
an absolute waiver of all civil liability in connection with the
Program, unless a warranty or assumption of liability accompanies a
copy of the Program in return for a fee.
END OF TERMS AND CONDITIONS
How to Apply These Terms to Your New Programs
If you develop a new program, and you want it to be of the greatest
possible use to the public, the best way to achieve this is to make it
free software which everyone can redistribute and change under these terms.
To do so, attach the following notices to the program. It is safest
to attach them to the start of each source file to most effectively
state the exclusion of warranty; and each file should have at least
the "copyright" line and a pointer to where the full notice is found.
<one line to give the program's name and a brief idea of what it does.>
Copyright (C) <year> <name of author>
This program is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation, either version 3 of the License, or
(at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program. If not, see <https://www.gnu.org/licenses/>.
Also add information on how to contact you by electronic and paper mail.
If the program does terminal interaction, make it output a short
notice like this when it starts in an interactive mode:
<program> Copyright (C) <year> <name of author>
This program comes with ABSOLUTELY NO WARRANTY; for details type `show w'.
This is free software, and you are welcome to redistribute it
under certain conditions; type `show c' for details.
The hypothetical commands `show w' and `show c' should show the appropriate
parts of the General Public License. Of course, your program's commands
might be different; for a GUI interface, you would use an "about box".
You should also get your employer (if you work as a programmer) or school,
if any, to sign a "copyright disclaimer" for the program, if necessary.
For more information on this, and how to apply and follow the GNU GPL, see
<https://www.gnu.org/licenses/>.
The GNU General Public License does not permit incorporating your program
into proprietary programs. If your program is a subroutine library, you
may consider it more useful to permit linking proprietary applications with
the library. If this is what you want to do, use the GNU Lesser General
Public License instead of this License. But first, please read
<https://www.gnu.org/licenses/why-not-lgpl.html>.
Pictures/BMEL.jpg
BIN
View File
Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 7.2 KiB

Pictures/BMEL_dark.png
BIN
View File
Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 17 KiB

Pictures/DBFZ.png
BIN
View File
Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 14 KiB

Pictures/DBFZ_dark.png
BIN
View File
Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 15 KiB

Pictures/MRI.png
BIN
View File
Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 8.7 KiB

Pictures/gradio_app.png
BIN
View File
Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 131 KiB

Pictures/kida.png
BIN
View File
Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 16 KiB

Pictures/kida_dark.png
BIN
View File
Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 17 KiB

Pictures/pipeline.png
BIN
View File
Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 901 KiB

Pictures/pipeline_light.png
BIN
View File
Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 1.1 MiB

README.md
+199 -1
View File
@@ -1 +1,199 @@
# transcriptor
# `ScrAIbe: Streamlined Conversation Recording with Automated Intelligence Based Environment`
`ScrAIbe` is a state-of-the-art, [PyTorch](https://pytorch.org/) based multilingual speech-to-text framework to generate fully automated transcriptions.
Beyond transcription, ScrAIbe supports advanced functions, such as speaker diarization and speaker recognition.
Designed as a comprehensive AI toolkit, it uses multiple AI models:
- [whisper](https://github.com/openai/whisper): A general-purpose speech recognition model.
- [payannote-audio](https://github.com/pyannote/pyannote-audio): An open-source toolkit for speaker diarization.
The framework utilizes a PyanNet-inspired pipeline, with the `Pyannote` library for speaker diarization and `VoxCeleb` for speaker embedding.
During post-diarization, each audio segment is processed by the OpenAI `Whisper` model, in a transformer encoder-decoder structure. Initially, a CNN mitigates noise and enhances speech. Before transcription, `VoxLingua` identifies the language segment, facilitating Whisper's role in both transcription and text translation.
The following graphic illustrates the whole pipeline:
![Pipeline](Pictures/pipeline.png#gh-dark-mode-only)
![Pipeline](Pictures/pipeline_light.png#gh-light-mode-only)
## Install `ScrAIbe` :
The following command will pull and install the latest commit from this repository, along with its Python dependencies.
pip install scraibe
- **Python version**: Python 3.8
- **PyTorch version**: Python 1.11.0
- **CUDA version**: Cuda-toolkit 11.3.1
Important: For the `Pyannote` model, you need to be granted access to Hugging Face.
Check the [Pyannote model page](https://huggingface.co/pyannote/speaker-diarization) to get access to the model.
Additionally, you need to generate a [Hugging Face token](https://huggingface.co/docs/hub/security-tokens).
## Usage
We've developed ScrAIbe with several access points to cater to diverse user needs.
### Python usage
It enables full control over the functionalities as well as process customization.
```python
from scraibe import Scraibe
model = Scraibe(use_auth_token = "hf_yourhftoken")
text = model.autotranscribe("audio.wav")
print(f"Transcription: \n{text}")
```
The `Scraibe` Class is taking care of the models being properly loaded. Therefore, you can choose the other [whisper](https://github.com/openai/whisper/blob/main/model-card.md) models using the `whisper_model` keyword.
You can also change the `pyannote` diarization model using the `dia_model` keyword.
As input, `autoranscribe` accepts every format which is compatible with [ffmgeg](https://ffmpeg.org/ffmpeg-formats.html). Examples therefore are `.mp4 .mp3 .wav .ogg .flac` and many more.
To further control the pipeline of `ScrAIbe` you can parse almost any keyword you also cloud parsed towards `whisper` or `pyannote` if you need more option, try to check out the documentations tows two Frameworks, you might have a good chance that these keywords will work here as well.
Here's are some examples regarding the `diarization` (which relies on the `pyannote` pipeline):
- `num_speakers` Number of speakers in the audio file
- `min_speakers` Minimal Number of speakers in the audio file
- `max_speakers` maximal Number of speakers in the audio file
Then there are arguments about the transcription process, which uses the "whisper" model.
- `language` Specify the language ([list to supported languages](https://github.com/openai/whisper/blob/main/language-breakdown.svg))
- `task` can be just `transcribe` or `translate`. If `translate` is selected, the transcribed audio will be translated to English.
For example:
```
text = model.autotranscribe("audio.wav", language="german", num_speakers = 2)
```
`Scraibe` also contains the option to just do a transcription
```python
transcription = model.transcribe("audio.wav")
```
or just do a diarization:
```python
diarization = model.diarize("audio.wav")
```
### Command-line usage
Next to the Pyhton interface, you can also run ScrAIbe using the command-line interface:
scraibe -f "audio.wav" --hf-token "hf_yourhftoken" --language "german" --num_speakers 2
For the full list of options, run:
scraibe -h
### Gradio App
The Gradio App is a user-friendly interface for ScrAIbe. It enables you to run the model without any coding knowledge. Therefore, you can run the app in your browser and upload your audio file, or you can make the Framework avail on your network and run it on your local machine.
#### Running the Gradio App on your local machine
To run the Gradio App on your local machine, just use the following command:
```
scraibe --start_server --port 7860 --hf_token hf_yourhftoken
```
- `--start_server`: Command to start the Gradio App.
- `--port`: Flag for connecting the container internal port to the port on your local machine.
- `--hf_token`: Flag for entering your personal HuggingFace token in the container.
When the app is running, it will show you at which address you can access it.
The default address is: http://127.0.0.1:7860 or http://0.0.0.0:7860
After the app is running, you can upload your audio file and select the desired options.
An example is shown below:
![Gradio App](Pictures/gradio_app.png)
### Running a Docker container
Another option to run ScrAIbe is to use a Docker container. This option is especially useful if you want to run the model on a server or if you would like to use the GPU without dealing with CUDA.
After you have installed Docker, you can execute the following commands in the terminal.
First, you need to build the Docker image. Therefore, you need to enter your HuggingFace token and the image name.
```
docker build . --build-arg="hf_token=[enter your HuggingFace token] " -t scraibe
```
After the image is built, you can run the container with the following command:
```
sudo docker run -it -p 7860:7860 --name [container name][image name] --hf_token [enter your HuggingFace token] --start_server
```
- `-p`: Flag for connecting the container internal port to the port on your local machine.
- `--hf_token`: Flag for entering your personal HuggingFace token in the container.
- `--start_server`: Command to start the Gradio App.
Inside the container, the `cli` is used. Therefore, you can use the same commands as in the command-line interface.
#### Enabling GPU usage
To use the GPU, ensure your Docker installation supports GPU usage.
For further information, check: https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/install-guide.html#docker
To enable GPU usage, you need to add the following flag to the `docker run` command:
```
docker run -it -p 7860:7860 --gpus 'all,capabilities=utility' --name [container name][image name] --hf_token [enter your HuggingFace token] --start_server
```
For further guidance, check: https://blog.roboflow.com/use-the-gpu-in-docker/
## Documentation
For further insights, check the [documentation page]().
## Contributions
We are happy to have any interest in contributing and about feedback: In order to do that, create an issue with your feedback or feel free to contact us.
## Roadmap
The following milestones are planned for further releases of ScrAIbe:
- Model quantization
Quantization to empower memory and computational efficiency.
- Model fine-tuning
In order to be able to cover a variety of linguistic phenomena.
For example, currently ScrAIbe is able to transcribe word by word, but ignores filler words or speech pauses.
These phenomena can be addressed by fine-tuning with the corresponding data.
- Implementation of LLMs
One example is the implementation of a summarization or extraction model, which enables ScrAIbe to automatically summarize or retrieve the key information out of a generated transcription, which could be the minutes of a meeting.
- Executable for Windows
## Contact
For queries contact [Jacob Schmieder](Jacob.Schmieder@dbfz.de)
## License
ScrAIbe is licensed under GNU General Public License.
## Acknowledgments
Special thanks go to the KIDA project and the BMEL (Bundesministerium für Ernährung und Landwirtschaft), especially to the AI Consultancy Team.
![KIDA](Pictures/kida_dark.png#gh-dark-mode-only) &nbsp; ![BMEL](Pictures/BMEL_dark.png#gh-dark-mode-only) &nbsp;&nbsp;&nbsp;&nbsp; ![DBFZ](Pictures/DBFZ_dark.png#gh-dark-mode-only) &nbsp; &nbsp;&nbsp;&nbsp; ![MRI](Pictures/MRI.png#gh-dark-mode-only)
![KIDA](Pictures/kida.png#gh-light-mode-only) &nbsp; ![BMEL](Pictures/BMEL.jpg#gh-light-mode-only) &nbsp;&nbsp;&nbsp;&nbsp; ![DBFZ](Pictures/DBFZ.png#gh-light-mode-only) &nbsp; &nbsp;&nbsp;&nbsp; ![MRI](Pictures/MRI.png#gh-light-mode-only)
autotranscript/.pyannotetoken
View File
autotranscript/__init__.py
-4
View File
@@ -1,4 +0,0 @@
from autotranscript.__main__ import *
from autotranscript.version import get_version as _get_version
__version__ = _get_version()
autotranscript/__main__.py
-497
View File
@@ -1,497 +0,0 @@
import whisper
from time import time, sleep
import os
import glob
import re
import shutil
import sys
from tqdm import tqdm
from typing import Union
from pydub import AudioSegment
from pyannote.audio import Pipeline
class AudioProcessor:
def __init__(self, audio_file:str):
self.audio_file_path = audio_file
self.audio_file = AudioSegment.from_file(audio_file, format=audio_file.split('.')[-1])
self.audiofilename = audio_file.split('/')[-1][:-4]
self.coreaudiofile = audio_file.split('/')[-1][:-4]
self.audiofilefolder = os.path.dirname(audio_file)
self.audio_file_type = audio_file.split('.')[-1]
def convert_audio(self, savefolder: str = "", savename: str = "", type: str = "wav", remove_orginal: bool = True):
"""
Convert video file or other audio files to mp3 file, ensures that the audio file is in the correct format for the
Whisper model
:param file: path to audio or video file
:param remove_orginal: remove original file
:return: mp3 file path
"""
print(f'Converting {self.audiofilename} to .{type} file')
if savefolder == "":
savefolder = self.audiofilefolder
if savename == "":
savename = self.coreaudiofile + f'.{type}'
else:
savename = savename + f'.{type}'
savepath = os.path.join(savefolder, savename)
self.audio_file.export(savepath, format=type)
print(f'Converted {self.audiofilename} to {type}')
if remove_orginal:
os.remove(self.audio_file_path)
print(f'File {self.audio_file_path} removed')
self.audio_file_path = savepath
self.audio_file = AudioSegment.from_file(savepath, format=type)
return self
def to_mp3(self, savefolder: str = "", savename: str = "", remove_orginal: bool = True):
"""
Convert audio file to mp3 file
:param file: audio file
:param remove_orginal: remove original file
:return: mp3 file path
"""
return self.convert_audio(savefolder = savefolder, savename = savename, type="mp3", remove_orginal=remove_orginal)
def to_wav(self, savefolder: str = "", savename: str = "", remove_orginal: bool = True):
"""
Convert audio file to wav file
:param file: audio file
:param remove_orginal: remove original file
:return: wav file path
"""
return self.convert_audio(savefolder = savefolder, savename = savename,type="wav", remove_orginal=remove_orginal)
def slower_mp3(self, savefolder: str = "", savename: str = "", speed: float = 0.75, type: str = "mp3"):
"""
Slow down mp3 file
:param file: mp3 file
:param speed: speed
:return: None
"""
if savefolder == "":
savefolder = self.audiofilefolder
else:
savefolder = savefolder
sound = self.audio_file
slow_sound = sound._spawn(sound.raw_data, overrides={
"frame_rate": int(sound.frame_rate * speed)
})
speedstr = str(speed).replace('.', '')
file_out = self.coreaudiofile + f'_{speedstr}.{type}'
save_path = os.path.join(savefolder, file_out)
slow_sound.export(save_path, format=type)
return slow_sound
class WhisperTranscription:
def __init__(self, audio_file: str , model, language: str = "German"):
self.audio_file = audio_file
self.model = model
self.language = language
def transcribe(self, language:str = "German"):
"""
Transcribe audio file
language: language of the audio file
:return: transcript as string
"""
audiofilename = self.audio_file.split('/')[-1]
#print(f'Start transcribing Audio file: {audiofilename}')
_stime = time()
result = self.model.transcribe(self.audio_file, language=self.language)
#print(f'Transcription finished in {time() - _stime} seconds')
self.transcript = result
return result["text"]
def save_transcript(self, transcript:str = "", savefolder : str = "", savename: str = ""):
"""
Save transcript to file
:param transcript: transcript as string
:param savefolder: folder to save transcript
:param savename: name of the transcript file
:return: None
"""
if savefolder == "":
savefolder = os.path.dirname(self.audio_file)
else:
savefolder = savefolder
if savename == "":
savename = self.audio_file.split('/')[-1][:-4] + '.txt'
else:
savename = savename
if transcript == "":
transcript = self.transcript["text"]
savepath = os.path.join(savefolder, savename)
with open(savepath, 'w') as f:
f.write(transcript)
print(f'Transcript saved to {savepath}')
class Diarisation(AudioProcessor):
def __init__(self, audio_file: str, model,**kwargs):
super().__init__(audio_file=audio_file)
self.model = model
def diarization(self, *args, **kwargs):
if "num_speakers" in kwargs:
num_speakers = kwargs['num_speakers']
kwargs.pop('num_speakers')
else:
num_speakers = 2
audiofilename = self.coreaudiofile
print(f'Start diarization of audio file: {self.audiofilename}')
_stime = time()
diarization = self.model(self.audio_file_path, num_speakers=num_speakers)
print(f'Diarization finished in {time() - _stime} seconds')
self.diarization = diarization
return diarization
def format_diarization_output(self, *args, **kwargs):
"""
Format diarization output to a list of tuples
:param args:
:param kwargs:
:return: dict with speaker names as keys and list of tuples as values and list of different speakers
"""
diarization_output = {"speakers": [], "segments": []}
if not hasattr(self, 'diarization'):
# ensure diarization is run before formatting
self.diarization = self.diarization()
for segment, _, speaker in self.diarization.itertracks(yield_label=True):
diarization_output["speakers"].append(speaker)
diarization_output["segments"].append(segment)
normalized_output = []
index_start_speaker = 0
index_end_speaker = 0
current_speaker = str()
for i, speaker in enumerate(diarization_output["speakers"]):
if i == 0:
current_speaker = speaker
if speaker != current_speaker:
index_end_speaker = i - 1
normalized_output.append([index_start_speaker, index_end_speaker, current_speaker])
index_start_speaker = i
current_speaker = speaker
if i == len(diarization_output["speakers"]) - 1:
index_end_speaker = i
normalized_output.append([index_start_speaker, index_end_speaker, current_speaker])
self.normalized_output = normalized_output
self.diarization_output = diarization_output
return diarization_output,normalized_output
def create_temporary_wav(self,savefolder: str = "", savename: str = "", *args, **kwargs):
"""
Create temporary wav file for diarization
:param savefolder: folder to save the temporary wav file
:param savename: name of the temporary wav file prefix
:param audiofile: audio file
:return: temporary wav file
"""
if savefolder == "":
folder = '.temp'
if not os.path.exists(folder):
os.makedirs(folder)
else:
folder = savefolder
folder = os.path.realpath(folder)
if savename == "":
savename = self.coreaudiofile + '.wav'
else:
savename = savename
if not os.path.exists(folder):
os.makedirs(folder)
if not hasattr(self, 'normalized_output') or not hasattr(self, 'diarization_output'):
self.format_diarization_output()
speaker = set(self.diarization_output["speakers"])
num_speak_iter = [0 for _ in range(len(speaker))]
for count, outp in enumerate(self.normalized_output):
start = self.diarization_output["segments"][outp[0]].start
end = self.diarization_output["segments"][outp[1]].end
print("start: ", start)
print("end: ", end)
start_milliseconds = start * 1000
end_milliseconds = end * 1000
print("start_milliseconds: ", start_milliseconds)
print("end_milliseconds: ", end_milliseconds)
print("cut audio")
cut_audio = self.audio_file[start_milliseconds:end_milliseconds]
print("save audio")
print(f".temp/{count}_speaker_" + str(outp[2]) + ".wav")
cut_audio.export(f".temp/{count}_speaker_" + str(outp[2]) + ".wav", format="wav")
return os.path.realpath(folder)
def __repr__(self):
return f"Diarization(audiofile={self.audiofile}, model={self.model}, language={self.language})"
def __str__(self):
return f"Diarization(audiofile={self.audiofile}, model={self.model}, language={self.language})"
class AutoTranscribe:
def __init__(self, audiofile: Union[str, bool, list] = None,
model: str = "medium",
language: str = "German",
diarisation: bool = False,
audioinput: str = "audiofiles",
transcriptionout: str = "transcriptions",
*args, **kwargs):
"""
AutoTranscribe
:param audiofile: audio file or list of audio files to transcribe
:param model: model name (default: medium)
:param language: language (default: German)
:param diarisation: diarisation (default: False)
"""
if audiofile is None:
audiofile = os.listdir(audioinput) # get all audio files in audioinput folder
audiofile = [os.path.realpath(os.path.join(audioinput, file)) for file in audiofile]# add path to audio files
self.audiofile = audiofile
self.language = language
self.diarisation = diarisation
if diarisation:
print("Diarisation is enabled")
print("Load Diarisation model")
self.diarisation_model = Pipeline.from_pretrained("pyannote/speaker-diarization",
use_auth_token = self._get_token())
print("Load Diarisation model done")
print(f"Load Whisper model {model}")
self.model = whisper.load_model(model)
print(f"Load Whisper model {model} done")
self.currentpath, \
self.audiopath, \
self.transcriptionpath, \
self.audiofiles = self.create_folder_structure(audioinput, transcriptionout) # create folder structure
def transcribe(self, *args, **kwargs):
if isinstance(self.audiofile, str):
for i in range(len(self.audiofiles)):
if self.audiofile in self.audiofiles[i]:
self.audiofile = [self.audiofiles[i]]
break
audiolist = self.audiofile
elif isinstance(self.audiofile, list):
audiolist = self.audiofile
else:
audiolist = self.audiofiles
if not set(audiolist).issubset(set(self.audiofiles)):
raise ValueError(f"Audio file {self.audiofile} not found in {self.audiopath}")
for audiofile in audiolist:
_start = time()
if not "/" in audiofile:
audiofile = os.path.join(self.audiopath, audiofile)
if not self.check_if_already_transcribed (audiofile):
audio = AudioProcessor(audiofile)
if not audiofile.endswith('wav'):
audio = audio.to_wav()
self.audiofile = audio.audio_file_path
audiofile = audio.audio_file_path
if "speed" in kwargs:
speed = kwargs['speed']
kwargs.pop('speed')
print('Creating slower version of the audio file with speed {}'.format(speed))
slower_audio = os.path.join(self.transcriptionpath, 'slower_version')
if not os.path.exists(slower_audio):
os.makedirs(slower_audio)
audio.slower_mp3(savefolder=slower_audio,speed=speed)
if not self.diarisation:
WhisperTranscription(audiofile, self.model, self.language
).save_transcript(savefolder = self.transcriptionpath)
else:
print("Start diarisation")
dia = Diarisation(audiofile, self.diarisation_model)
if 'num_speakers' in kwargs:
num_speakers = kwargs['num_speakers']
kwargs.pop('num_speakers')
dia.diarization(num_speakers=num_speakers)
else:
dia.diarization()
temppath = dia.create_temporary_wav()
temppath_dict, _ = dia.format_diarization_output()
speakers = list(set(temppath_dict["speakers"]))
fstring = "\\begin{drama}"
for speaker in speakers:
speaker = speaker.replace("SPEAKER_", "")
fstring += "\n\t\Character{S"+ str(speaker) + "}{S" + str(speaker) + "}"
files = glob.glob(temppath + "/*.wav")
# Sort files according to the digits included in the filename
files = sorted(files, key=lambda x: float(re.findall("(\d+)", x)[0]))
for file in tqdm(files):
Whisper = WhisperTranscription(file, self.model, self.language).transcribe()
for s in speakers:
if s in file:
s = s.replace("SPEAKER_", "")
fstring += f"\n\S{s}speaks: \n {Whisper}"
fstring += "\n\end{drama}"
print(fstring)
with open(os.path.join(self.transcriptionpath,
os.path.basename(audiofile).split('.')[0] + '.tex'), 'w') as f:
f.write(fstring)
print("Remove temporary files")
shutil.rmtree(temppath)
print(f"Transcription of {audiofile} done in total of {time() - _start} seconds")
def create_folder_structure(self, audiopath: str, transcriptionout: str):
"""
Create folder structure for audio and transcription files
:return: currentpath, audiopath, transcriptionpath, audiofiles
"""
currentpath = os.path.dirname(sys.argv[0]) # get executable path
if not os.path.exists(os.path.join(currentpath, audiopath)):
print('Creating audiofiles folder')
os.makedirs(os.path.join(currentpath, audiopath))
if not os.path.exists(os.path.join(currentpath, transcriptionout)):
print('Creating transcription folder')
os.makedirs(os.path.join(currentpath, transcriptionout))
audiopath = os.path.join(currentpath, audiopath) # path to audio files
transcriptionpath = os.path.join(currentpath, transcriptionout) # path to transcription files
_audiofiles = os.listdir(audiopath) # list of audio files
audiofiles = []
for i in _audiofiles:
audiofiles.append(os.path.join(audiopath, i))
return currentpath, audiopath, transcriptionpath, audiofiles
def check_if_already_transcribed (self, filename: str):
"""
Check if all audio files are already transcribed
:param filename: audio file name
:return: bool
"""
purefilename = filename.split('/')[-1][:-4]
_files = os.listdir(self.transcriptionpath)
for i,f in enumerate(_files):
_files[i] = f[:-4]
if purefilename in _files:
print(f'File {purefilename[:-4]} already transcribed')
return True
else:
return False
@classmethod
def _get_token(self):
# check ig .pyannotetoken.txt exists
path = os.path.join(os.path.dirname(os.path.realpath(__file__)), '.pyannotetoken')
if os.path.exists(path):
with open(path, 'r') as f:
token = f.read()
else:
raise ValueError('No token found. Please create a token at https://huggingface.co/settings/token'
' and save it in a file called .pyannotetoken.txt')
return token
def __repr__(self):
return f"AutoTranscribe(audiofile={self.audiofile}, model={self.model}, language={self.language}, diarisation={self.diarisation})"
def __call__(self, *args, **kwargs):
return self.transcribe(*args, **kwargs)
autotranscript/__pycache__/__init__.cpython-39.pyc
BIN
View File
Binary file not shown.
autotranscript/__pycache__/__main__.cpython-39.pyc
BIN
View File
Binary file not shown.
environment.yml
+59 -10
View File
@@ -1,4 +1,3 @@
name: whisper
channels:
- pytorch
- defaults
@@ -11,7 +10,7 @@ dependencies:
- ca-certificates=2023.05.30=h06a4308_0
- certifi=2023.5.7=py39h06a4308_0
- cffi=1.15.1=py39h5eee18b_3
- cryptography=39.0.1=py39h9ce1e76_0
- cryptography=39.0.1=py39h9ce1e76_2
- cudatoolkit=11.3.1=h2bc3f7f_2
- ffmpeg=4.2.2=h20bf706_0
- flit-core=3.8.0=py39h06a4308_0
@@ -51,36 +50,40 @@ dependencies:
- numpy=1.23.5=py39h14f4228_0
- numpy-base=1.23.5=py39h31eccc5_0
- openh264=2.1.1=h4ff587b_0
- openssl=1.1.1t=h7f8727e_0
- openssl=3.0.9=h7f8727e_0
- pillow=9.4.0=py39h6a678d5_0
- pip=23.0.1=py39h06a4308_0
- pycparser=2.21=pyhd3eb1b0_0
- pyopenssl=23.0.0=py39h06a4308_0
- pysocks=1.7.1=py39h06a4308_0
- python=3.9.16=h7a1cb2a_2
- python=3.9.16=h955ad1f_3
- pytorch=1.11.0=py3.9_cuda11.3_cudnn8.2.0_0
- pytorch-mutex=1.0=cuda
- readline=8.2=h5eee18b_0
- requests=2.28.1=py39h06a4308_1
- setuptools=65.6.3=py39h06a4308_0
- six=1.16.0=pyhd3eb1b0_1
- sqlite=3.41.1=h5eee18b_0
- sqlite=3.41.2=h5eee18b_0
- tk=8.6.12=h1ccaba5_0
- torchaudio=0.11.0=py39_cu113
- torchvision=0.12.0=py39_cu113
- typing_extensions=4.4.0=py39h06a4308_0
- tzdata=2023c=h04d1e81_0
- wheel=0.38.4=py39h06a4308_0
- x264=1!157.20191217=h7b6447c_0
- xz=5.2.10=h5eee18b_1
- xz=5.4.2=h5eee18b_0
- zlib=1.2.13=h5eee18b_0
- zstd=1.5.4=hc292b87_0
- pip:
- absl-py==1.3.0
- aiofiles==23.1.0
- aiohttp==3.8.3
- aiosignal==1.3.1
- alembic==1.9.1
- altair==5.0.1
- annotated-types==0.5.0
- ansi2html==1.8.0
- antlr4-python3-runtime==4.9.3
- anyio==3.7.1
- appdirs==1.4.4
- asteroid-filterbanks==0.4.0
- async-timeout==4.0.2
@@ -100,48 +103,76 @@ dependencies:
- commonmark==0.9.1
- contourpy==1.0.6
- cycler==0.11.0
- dash==2.12.1
- dash-core-components==2.0.0
- dash-html-components==2.0.0
- dash-table==5.0.0
- decorator==4.4.2
- docopt==0.6.2
- einops==0.3.2
- exceptiongroup==1.1.1
- fastapi==0.100.0
- ffmpeg-python==0.2.0
- ffmpy==0.3.0
- filelock==3.8.0
- flask==2.2.5
- fonttools==4.38.0
- frozenlist==1.3.3
- fsspec==2022.11.0
- future==0.18.2
- google-auth==2.15.0
- google-auth-oauthlib==0.4.6
- gradio==3.36.1
- gradio-client==0.2.7
- greenlet==2.0.1
- grpcio==1.51.1
- h11==0.14.0
- hmmlearn==0.2.8
- huggingface-hub==0.11.0
- httpcore==0.17.3
- httpx==0.24.1
- huggingface-hub==0.16.4
- humanize==4.7.0
- hyperpyyaml==1.1.0
- imageio==2.23.0
- imageio-ffmpeg==0.4.7
- importlib-metadata==4.13.0
- importlib-resources==5.12.0
- iniconfig==2.0.0
- itsdangerous==2.1.2
- jinja2==3.1.2
- joblib==1.2.0
- jsonschema==4.18.0
- jsonschema-specifications==2023.6.1
- julius==0.2.7
- kiwisolver==1.4.4
- librosa==0.9.2
- linkify-it-py==2.0.2
- lit==16.0.5.post0
- llvmlite==0.39.1
- mako==1.2.4
- markdown==3.4.1
- markdown-it-py==2.2.0
- markupsafe==2.1.1
- matplotlib==3.6.2
- matplotlib==3.7.1
- mdit-py-plugins==0.3.3
- mdurl==0.1.2
- more-itertools==9.0.0
- moviepy==1.0.3
- mpmath==1.2.1
- multidict==6.0.4
- nest-asyncio==1.5.7
- networkx==2.8.8
- numba==0.56.4
- oauthlib==3.2.2
- omegaconf==2.3.0
- openai-whisper==20230314
- optuna==3.0.5
- orjson==3.9.2
- packaging==21.3
- pandas==1.5.2
- pbr==5.11.0
- plotly==5.15.0
- pluggy==1.0.0
- pooch==1.6.0
- prettytable==3.5.0
- primepy==1.3
@@ -154,23 +185,32 @@ dependencies:
- pyannote-pipeline==2.3
- pyasn1==0.4.8
- pyasn1-modules==0.2.8
- pydantic==2.0.2
- pydantic-core==2.1.2
- pydeprecate==0.3.2
- pydub==0.25.1
- pygments==2.13.0
- pyparsing==3.0.9
- pyperclip==1.8.2
- pytest==7.3.1
- python-dateutil==2.8.2
- python-multipart==0.0.6
- pytorch-lightning==1.6.5
- pytorch-metric-learning==1.6.3
- pytz==2022.7
- pyyaml==6.0
- qtfaststart==1.8
- referencing==0.29.1
- regex==2022.10.31
- requests-oauthlib==1.3.1
- resampy==0.4.2
- retrying==1.3.4
- rich==12.6.0
- rpds-py==0.8.10
- rsa==4.9
- ruamel-yaml==0.17.21
- ruamel-yaml-clib==0.2.7
- ruff==0.0.272
- scikit-learn==1.2.0
- scipy==1.8.1
- semantic-version==2.10.0
@@ -180,19 +220,24 @@ dependencies:
- shellingham==1.5.0
- simplejson==3.18.0
- singledispatchmethod==1.0
- sniffio==1.3.0
- sortedcontainers==2.4.0
- soundfile==0.10.3.post1
- speechbrain==0.5.13
- speechbrain==0.5.14
- sqlalchemy==1.4.45
- starlette==0.27.0
- stevedore==4.1.1
- sympy==1.11.1
- tabulate==0.9.0
- tenacity==8.2.2
- tensorboard==2.11.0
- tensorboard-data-server==0.6.1
- tensorboard-plugin-wit==1.8.1
- threadpoolctl==3.1.0
- tiktoken==0.3.1
- tokenizers==0.13.2
- tomli==2.0.1
- toolz==0.12.0
- torch-audiomentations==0.11.0
- torch-pitch-shift==1.2.2
- torchmetrics==0.11.0
@@ -200,8 +245,12 @@ dependencies:
- transformers==4.24.0
- triton==2.0.0
- typer==0.7.0
- typing-extensions==4.7.1
- uc-micro-py==1.0.2
- urllib3==1.26.12
- uvicorn==0.22.0
- wcwidth==0.2.5
- websockets==11.0.3
- werkzeug==2.2.2
- yarl==1.8.2
- zipp==3.11.0
requirements.txt
+24 -151
View File
@@ -1,152 +1,25 @@
absl-py==1.3.0
aiohttp==3.8.3
aiosignal==1.3.1
alembic==1.9.1
antlr4-python3-runtime==4.9.3
appdirs==1.4.4
asteroid-filterbanks==0.4.0
async-timeout==4.0.2
attrs==22.2.0
audioread==3.0.0
autopage==0.5.1
backports.cached-property==1.0.2
brotlipy==0.7.0
cachetools==5.2.0
certifi==2023.5.7
cffi==1.15.1
charset-normalizer==2.1.1
click==8.1.3
cliff==4.1.0
cmaes==0.9.0
cmake==3.26.4
cmd2==2.4.2
colorama==0.4.6
colorlog==6.7.0
commonmark==0.9.1
contourpy==1.0.6
cryptography==39.0.1
cycler==0.11.0
decorator==4.4.2
docopt==0.6.2
einops==0.3.2
ffmpeg-python==0.2.0
filelock==3.8.0
flit_core==3.8.0
fonttools==4.38.0
frozenlist==1.3.3
fsspec==2022.11.0
future==0.18.2
google-auth==2.15.0
google-auth-oauthlib==0.4.6
greenlet==2.0.1
grpcio==1.51.1
hmmlearn==0.2.8
huggingface-hub==0.11.0
HyperPyYAML==1.1.0
idna==3.4
imageio==2.23.0
imageio-ffmpeg==0.4.7
importlib-metadata==4.13.0
joblib==1.2.0
julius==0.2.7
kiwisolver==1.4.4
librosa==0.9.2
lit==16.0.5.post0
llvmlite==0.39.1
Mako==1.2.4
Markdown==3.4.1
MarkupSafe==2.1.1
matplotlib==3.6.2
mkl-fft==1.3.1
mkl-random==1.2.2
mkl-service==2.4.0
more-itertools==9.0.0
moviepy==1.0.3
mpmath==1.2.1
multidict==6.0.4
networkx==2.8.8
numba==0.56.4
numpy==1.23.5
oauthlib==3.2.2
omegaconf==2.3.0
openai-whisper==20230314
optuna==3.0.5
packaging==21.3
pandas==1.5.2
pbr==5.11.0
Pillow==9.4.0
pip==23.0.1
pooch==1.6.0
prettytable==3.5.0
primePy==1.3
proglog==0.1.10
protobuf==3.20.1
pyannote.audio==2.1.1
pyannote.core==4.5
pyannote.database==4.1.3
pyannote.metrics==3.2.1
pyannote.pipeline==2.3
pyasn1==0.4.8
pyasn1-modules==0.2.8
pycparser==2.21
pyDeprecate==0.3.2
pydub==0.25.1
Pygments==2.13.0
pyOpenSSL==23.0.0
pyparsing==3.0.9
pyperclip==1.8.2
PySocks==1.7.1
python-dateutil==2.8.2
pytorch-lightning==1.6.5
pytorch-metric-learning==1.6.3
pytz==2022.7
PyYAML==6.0
regex==2022.10.31
requests==2.28.1
requests-oauthlib==1.3.1
resampy==0.4.2
rich==12.6.0
rsa==4.9
ruamel.yaml==0.17.21
ruamel.yaml.clib==0.2.7
scikit-learn==1.2.0
scipy==1.8.1
semantic-version==2.10.0
semver==2.13.0
sentencepiece==0.1.97
setuptools==65.6.3
setuptools-rust==1.5.2
shellingham==1.5.0
simplejson==3.18.0
singledispatchmethod==1.0
six==1.16.0
sortedcontainers==2.4.0
SoundFile==0.10.3.post1
speechbrain==0.5.13
SQLAlchemy==1.4.45
stevedore==4.1.1
sympy==1.11.1
tabulate==0.9.0
tensorboard==2.11.0
tensorboard-data-server==0.6.1
tensorboard-plugin-wit==1.8.1
threadpoolctl==3.1.0
tiktoken==0.3.1
tokenizers==0.13.2
torch==1.11.0
torch-audiomentations==0.11.0
torch-pitch-shift==1.2.2
torchaudio==0.11.0
torchmetrics==0.11.0
torchvision==0.12.0
tqdm==4.65.0
transformers==4.24.0
triton==2.0.0
typer==0.7.0
typing_extensions==4.4.0
urllib3==1.26.15
wcwidth==0.2.5
Werkzeug==2.2.2
wheel==0.38.4
yarl==1.8.2
zipp==3.11.0
pyannote.audio~=2.1.1
pyannote.core~=4.5
pyannote.database~=4.1.3
pyannote.metrics~=3.2.1
pyannote.pipeline~=2.3
setuptools~=65.6.3
setuptools-rust~=1.5.2
tqdm>=4.65.0
gradio~=3.36.1
gradio-client~=0.2.7
# add pytorch to override the one installed by pyannote.audio
torch~=1.11.0
torchvision~=0.12.0
torchaudio~=0.11.0
#optional:
#sphinx~=5.0.2
scraibe/.pyannotetoken
+1
View File
@@ -0,0 +1 @@
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
scraibe/__init__.py
+15
View File
@@ -0,0 +1,15 @@
from .autotranscript import *
from .transcriber import *
from .audio import *
from .transcript_exporter import *
from .diarisation import *
from .version import get_version as _get_version
from .misc import *
from .app.gradio_app import *
from .app.qtfaststart import *
from .cli import *
__version__ = _get_version()
scraibe/app/Logo_KIDA_bmel_green.svg
+171
View File
File diff suppressed because one or more lines are too long
Side by Side

After

Width:  |  Height:  |  Size: 38 KiB

scraibe/app/__init__.py
+2
View File
@@ -0,0 +1,2 @@
from .qtfaststart import *
from .gradio_app import *
scraibe/app/gradio_app.py
+438
View File
@@ -0,0 +1,438 @@
"""
Gradio Audio Transcription App.
--------------------------------
This module provides an interface to transcribe audio files using the
Scraibe model. Users can either upload an audio file or record their speech
live for transcription. The application supports multiple languages and provides
options to specify the number of speakers and the language of the audio.
Attributes:
LANGUAGES (list): A list of supported languages for transcription.
Usage:
Run this script to start the Gradio web interface for audio transcription.
"""
"""
Gradio Audio Transcription App.
--------------------------------
This module provides an interface to transcribe audio files using the
Scraibe model. Users can either upload an audio file or record their speech
live for transcription. The application supports multiple languages and provides
options to specify the number of speakers and the language of the audio.
Attributes:
LANGUAGES (list): A list of supported languages for transcription.
Usage:
Run this script to start the Gradio web interface for audio transcription.
"""
import json
import os
from tkinter import CURRENT
import gradio as gr
from tqdm import tqdm
from scraibe import Scraibe, Transcript
theme = gr.themes.Soft(
primary_hue="green",
secondary_hue='orange',
neutral_hue="gray",
)
LANGUAGES = [
"Afrikaans", "Arabic", "Armenian", "Azerbaijani", "Belarusian",
"Bosnian", "Bulgarian", "Catalan", "Chinese", "Croatian",
"Czech", "Danish", "Dutch", "English", "Estonian",
"Finnish", "French", "Galician", "German", "Greek",
"Hebrew", "Hindi", "Hungarian", "Icelandic", "Indonesian",
"Italian", "Japanese", "Kannada", "Kazakh", "Korean",
"Latvian", "Lithuanian", "Macedonian", "Malay", "Marathi",
"Maori", "Nepali", "Norwegian", "Persian", "Polish",
"Portuguese", "Romanian", "Russian", "Serbian", "Slovak",
"Slovenian", "Spanish", "Swahili", "Swedish", "Tagalog",
"Tamil", "Thai", "Turkish", "Ukrainian", "Urdu",
"Vietnamese", "Welsh"
]
CURRENT_PATH = os.path.dirname(os.path.realpath(__file__))
class GradioTranscriptionInterface:
"""
Interface handling the interaction between Gradio UI and the Audio Transcription system.
"""
def __init__(self, model: Scraibe):
"""
Initializes the GradioTranscriptionInterface with a transcription model.
Args:
model (Scraibe): Model responsible for audio transcription tasks.
"""
self.model = model
def auto_transcribe(self, source,
num_speakers : int,
translation : bool,
language : str):
"""
Shortcut method for the Scraibe task.
Returns:
tuple: Transcribed text (str), JSON output (dict)
"""
kwargs = {
"num_speakers": num_speakers if num_speakers != 0 else None,
"language": language if language != "None" else None,
"task": 'translate' if translation else None
}
if isinstance(source, str):
try:
result = self.model.autotranscribe(source, **kwargs)
except ValueError:
raise gr.Error("Couldn't detect any speech in the provided audio. \
Please try again!")
return str(result), result.get_json()
elif isinstance(source, list):
source_names = [s.split("/")[-1] for s in source]
result = []
for s in tqdm(source, total=len(source),desc = "Transcribing audio files"):
try:
res = self.model.autotranscribe(s, **kwargs)
except ValueError:
_name = s.split("/")[-1]
res = f"NO TRANSCRIPT FOUND FOR {_name}"
gr.Warning(f"Couldn't detect any speech in {_name} will skip this file.")
result.append(res)
out = ''
out_dict = {}
for i, r in enumerate(result):
out += f"TRANSCRIPT {i} FOR ({source_names[i]}):\n\n"
out += str(r)
out += "\n\n"
if isinstance(r, str):
out_dict[source_names[i]] = r
else:
out_dict[source_names[i]] = r.get_dict()
return out, json.dumps(out_dict, indent=4)
else:
raise gr.Error("Please provide a valid audio file.")
def transcribe(self, source, translation, language):
"""
Shortcut method for the Transcribe task.
Returns:
str: Transcribed text.
"""
kwargs = {
"language": language if language != "None" else None,
"task": 'translate' if translation == "Yes" else None
}
if isinstance(source, str):
result = self.model.transcribe(source, **kwargs)
return str(result)
elif isinstance(source, list):
source_names = [s.split("/")[-1] for s in source]
result = []
for s in tqdm(source, total=len(source),desc = "Transcribing audio files"):
res = self.model.transcribe(s, **kwargs)
result.append(res)
out = ''
for i, res in enumerate(result):
out += f"TRANSCRIPT {i} FOR ({source_names[i]}):\n\n"
out += str(res)
out += "\n\n"
return out
else:
raise gr.Error("Please provide a valid audio file.")
def perform_diarisation(self, source, num_speakers):
"""
Shortcut method for the Diarisation task.
Returns:
str: JSON output of diarisation result.
"""
kwargs = {
"num_speakers": num_speakers if num_speakers != 0 else None,
}
if isinstance(source, str):
try:
result = self.model.diarization(source, **kwargs)
except ValueError:
raise gr.Error("Couldn't detect any speech in the provided audio. \
Please try again!")
return json.dumps(result, indent=2)
elif isinstance(source, list):
source_names = [s.split("/")[-1] for s in source]
result = []
for s in tqdm(source, total=len(source),desc = "Performing diarisation"):
try:
res = self.model.diarization(s, **kwargs)
except ValueError:
res = f"NO DIARISATION FOUND FOR {s}"
gr.Warning(f"Couldn't detect any speech in {s} will skip this file.")
result.append(res)
out = {}
for i, res in enumerate(result):
out[source_names[i]] = res
return json.dumps(out, indent=4)
else:
gr.Error("Please provide a valid audio file.")
####
# Gradio Interface
####
def gradio_Interface(model : Scraibe = None):
if model is None:
model = Scraibe()
pipe = GradioTranscriptionInterface(model)
def select_task(choice):
if choice == 'Auto Transcribe':
return (gr.update(visible = True),
gr.update(visible = True),
gr.update(visible = True))
elif choice == 'Transcribe':
return (gr.update(visible = False),
gr.update(visible = True),
gr.update(visible = True))
elif choice == 'Diarisation':
return (gr.update(visible = True),
gr.update(visible = False),
gr.update(visible = False))
def select_origin(choice):
if choice == "Upload Audio":
return (gr.update(visible = True),
gr.update(visible = False, value = None),
gr.update(visible = False, value = None),
gr.update(visible = False, value = None),
gr.update(visible = False, value = None))
elif choice == "Record Audio":
return (gr.update(visible = False, value = None),
gr.update(visible = True),
gr.update(visible = False, value = None),
gr.update(visible = False, value = None),
gr.update(visible = False, value = None))
elif choice == "Upload Video":
return (gr.update(visible = False, value = None),
gr.update(visible = False, value = None),
gr.update(visible = True),
gr.update(visible = False, value = None),
gr.update(visible = False, value = None))
elif choice == "Record Video":
return (gr.update(visible = False, value = None),
gr.update(visible = False, value = None),
gr.update(visible = False, value = None),
gr.update(visible = True),
gr.update(visible = False, value = None))
elif choice == "File or Files":
return (gr.update(visible = False, value = None),
gr.update(visible = False, value = None),
gr.update(visible = False, value = None),
gr.update(visible = False, value = None),
gr.update(visible = True))
def run_scribe(task,
num_speakers,
translate,
language,
audio1,
audio2,
video1,
video2,
file_in,
progress = gr.Progress(track_tqdm= True)):
# get *args which are not None
progress(0, desc='Starting task...')
source = audio1 or audio2 or video1 or video2 or file_in
if isinstance(source, list):
source = [s.name for s in source]
if len(source) == 1:
source = source[0]
if task == 'Auto Transcribe':
out_str , out_json = pipe.auto_transcribe(source = source,
num_speakers = num_speakers,
translation = translate,
language = language)
if isinstance(source, str):
return (gr.update(value = out_str, visible = True),
gr.update(value = out_json, visible = True),
gr.update(visible = True),
gr.update(visible = True))
else:
return (gr.update(value = out_str, visible = True),
gr.update(value = out_json, visible = True),
gr.update(visible = False),
gr.update(visible = False))
elif task == 'Transcribe':
out = pipe.transcribe(source = source,
translation = translate,
language = language)
return (gr.update(value = out, visible = True),
gr.update(value = None, visible = False),
gr.update(visible = False),
gr.update(visible = False))
elif task == 'Diarisation':
out = pipe.perform_diarisation(source = source,
num_speakers = num_speakers)
return (gr.update(value = None, visible = False),
gr.update(value = out, visible = True),
gr.update(visible = False),
gr.update(visible = False))
def annotate_output(annoation : str, out_json : dict):
# get *args which are not None
trans = Transcript.from_json(out_json)
trans = trans.annotate(*annoation.split(","))
return gr.update(value = str(trans)),gr.update(value = trans.get_json())
with gr.Blocks(theme=theme,title='ScrAIbe: Automatic Audio Transcription') as demo:
# Define components
hname = os.path.join(CURRENT_PATH, "header.html")
header = open(hname, "r").read()
gr.HTML(header, visible= True, show_label=False)
with gr.Row():
with gr.Column():
task = gr.Radio(["Auto Transcribe", "Transcribe", "Diarisation"], label="Task",
value= 'Auto Transcribe')
num_speakers = gr.Number(value=0, label= "Number of speakers (optional)",
info = "Number of speakers in the audio file. If you don't know,\
leave it at 0.", visible= True)
translate = gr.Checkbox(label="Translation", choices=[True, False], value = False,
info="Select 'Yes' to have the output translated into English.",
visible= True)
language = gr.Dropdown(LANGUAGES,
label="Language (optional)", value = "None",
info="Language of the audio file. If you don't know,\
leave it at None.", visible= True)
input = gr.Radio(["Upload Audio", "Record Audio", "Upload Video","Record Video"
,"File or Files"], label="Input Type", value="Upload Audio")
audio1 = gr.Audio(source="upload", type="filepath", label="Upload Audio",
interactive= True, visible= True)
audio2 = gr.Audio(source="microphone", label="Record Audio", type="filepath",
interactive= True, visible= False)
video1 = gr.Video(source="upload", type="filepath", label="Upload Video",
interactive= True, visible= False)
video2 = gr.Video(source="webcam", label="Record Video", type="filepath",
interactive= True, visible= False)
file_in = gr.Files(label="Upload File or Files", interactive= True, visible= False)
submit = gr.Button()
with gr.Column():
out_txt = gr.Textbox(label="Output",
visible= True, show_copy_button=True)
out_json = gr.JSON(label="JSON Output",
visible= False, show_copy_button=True)
annoation = gr.Textbox(label="Name your speaker's",
info= "Please provide a list of the speakers arranged \
in the order in which they appear in the input. Use comma ',' \
as a seperator. Be aware that the first name is given \
to SPEAKER_00 the second to SPEAKER_01 and so on.",
visible= False, interactive= True)
annotate = gr.Button(value="Annotate", visible= False, interactive= True)
# Define usage of components
input.change(fn=select_origin, inputs=[input],
outputs=[audio1, audio2, video1, video2, file_in])
task.change(fn=select_task, inputs=[task],
outputs=[num_speakers, translate, language])
translate.change(fn= lambda x : gr.update(value = x),
inputs=[translate], outputs=[translate])
num_speakers.change(fn= lambda x : gr.update(value = x),
inputs=[num_speakers], outputs=[num_speakers])
language.change(fn= lambda x : gr.update(value = x),
inputs=[language], outputs=[language])
submit.click(fn = run_scribe,
inputs=[task, num_speakers, translate, language, audio1,
audio2, video1, video2, file_in],
outputs=[out_txt, out_json, annoation, annotate])
annotate.click(fn = annotate_output, inputs=[annoation, out_json],
outputs=[out_txt, out_json])
return demo
if __name__ == "__main__":
gradio_Interface().queue().launch()
scraibe/app/header.html
+66
View File
@@ -0,0 +1,66 @@
<!-- Importing Cormorant Garamond font from Google Fonts -->
<link href="https://fonts.googleapis.com/css2?family=Cormorant+Garamond:wght@400;700&display=swap" rel="stylesheet">
<style>
.header-container {
display: flex;
align-items: center;
justify-content: center;
position: relative;
padding-top: 30px;
}
.logo-container {
position: absolute;
top: 50%;
right: 20px;
transform: translateY(-50%);
width: 300px;
}
.logo {
width: 100%;
height: auto;
}
h1 {
font-family: 'Cormorant Garamond', serif;
font-size: 50px !important; /* Increased font size */
font-weight: bold;
color: #50AF31;
margin: 0;
position: relative;
padding: 0.5em 0;
}
h1::before, h1::after {
content: "";
position: absolute;
height: 2px;
width: 80%;
background-color: #50AF31;
left: 10%;
}
h1::before {
top: 0.5em;
}
h1::after {
bottom: 0.5em;
}
p, h2 {
font-size: 16px;
margin: 10px 0;
line-height: 1.4;
}
</style>
<div class="header-container">
<h1>ScrAIbe</h1>
<div class="logo-container">
<a href="https://www.kida-bmel.de/"> <!-- Replace with your actual URL -->
<img src="file/Logo_KIDA_bmel_green.svg" alt="KIDA Logo" class="logo">
</a>
</div>
</div>
<div style="text-align: center; padding: 20px 10%;">
<p>
Upload, record, or provide a video with audio for transcription. Our toolkit is designed to transcribe content from multiple languages accurately. The integrated speaker diarisation feature identifies different speakers, ensuring a smooth transcription experience. For optimal results, indicate the number of speakers and the original language of the content.
</p>
<h2 style="font-weight: bold; color: #50AF31;">What would you like to do next?</h2>
</div>
scraibe/app/qtfaststart.py
+319
View File
@@ -0,0 +1,319 @@
"""
This file contains a modified version of qtfaststart by qtfaststart
https://github.com/danielgtaylor/qtfaststart/tree/master
All credit goes to the original author.
Copyright (C) 2008 - 2013 Daniel G. Taylor <dan@programmer-art.org>
Permission is hereby granted, free of charge, to any person obtaining a copy of this
software and associated documentation files (the "Software"),
to deal in the Software without restriction, including without limitation the rights to
use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the
Software, and to permit persons to whom the Software is furnished to do so,
subject to the following conditions:
The above copyright notice and this permission notice shall be included in all copies
or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED,
INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM,
DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
IN THE SOFTWARE.
"""
import logging
import os
import struct
import collections
import io
# define error classes
class FastStartException(Exception):
"""
Raised when something bad happens during processing.
"""
pass
class FastStartSetupError(FastStartException):
"""
Rasised when asked to process a file that does not need processing
"""
pass
class MalformedFileError(FastStartException):
"""
Raised when the input file is setup in an unexpected way
"""
pass
class UnsupportedFormatError(FastStartException):
"""
Raised when a movie file is recognized as a format not supported.
"""
pass
# define constants
CHUNK_SIZE = 8192
log = logging.getLogger("qtfaststart")
# Older versions of Python require this to be defined
if not hasattr(os, 'SEEK_CUR'):
os.SEEK_CUR = 1
Atom = collections.namedtuple('Atom', 'name position size')
def read_atom(datastream):
"""
Read an atom and return a tuple of (size, type) where size is the size
in bytes (including the 8 bytes already read) and type is a "fourcc"
like "ftyp" or "moov".
"""
size, type = struct.unpack(">L4s", datastream.read(8))
type = type.decode('ascii')
return size, type
def _read_atom_ex(datastream):
"""
Read an Atom from datastream
"""
pos = datastream.tell()
atom_size, atom_type = read_atom(datastream)
if atom_size == 1:
atom_size, = struct.unpack(">Q", datastream.read(8))
return Atom(atom_type, pos, atom_size)
def get_index(datastream):
"""
Return an index of top level atoms, their absolute byte-position in the
file and their size in a list:
index = [
("ftyp", 0, 24),
("moov", 25, 2658),
("free", 2683, 8),
...
]
The tuple elements will be in the order that they appear in the file.
"""
log.debug("Getting index of top level atoms...")
index = list(_read_atoms(datastream))
_ensure_valid_index(index)
return index
def _read_atoms(datastream):
"""
Read atoms until an error occurs
"""
while datastream:
try:
atom = _read_atom_ex(datastream)
log.debug("%s: %s" % (atom.name, atom.size))
except:
break
yield atom
if atom.size == 0:
if atom.name == "mdat":
# Some files may end in mdat with no size set, which generally
# means to seek to the end of the file. We can just stop indexing
# as no more entries will be found!
break
else:
# Weird, but just continue to try to find more atoms
continue
datastream.seek(atom.position + atom.size)
def _ensure_valid_index(index):
"""
Ensure the minimum viable atoms are present in the index.
Raise FastStartException if not.
"""
top_level_atoms = set([item.name for item in index])
for key in ["moov", "mdat"]:
if key not in top_level_atoms:
log.error("%s atom not found, is this a valid MOV/MP4 file?" % key)
raise FastStartException()
def find_atoms(size, datastream):
"""
Compatibilty interface for _find_atoms_ex
"""
fake_parent = Atom('fake', datastream.tell()-8, size+8)
for atom in _find_atoms_ex(fake_parent, datastream):
yield atom.name
def _find_atoms_ex(parent_atom, datastream):
"""
Yield either "stco" or "co64" Atoms from datastream.
datastream will be 8 bytes into the stco or co64 atom when the value
is yielded.
It is assumed that datastream will be at the end of the atom after
the value has been yielded and processed.
parent_atom is the parent atom, a 'moov' or other ancestor of CO
atoms in the datastream.
"""
stop = parent_atom.position + parent_atom.size
while datastream.tell() < stop:
try:
atom = _read_atom_ex(datastream)
except:
log.exception("Error reading next atom!")
raise FastStartException()
if atom.name in ["trak", "mdia", "minf", "stbl"]:
# Known ancestor atom of stco or co64, search within it!
for res in _find_atoms_ex(atom, datastream):
yield res
elif atom.name in ["stco", "co64"]:
yield atom
else:
# Ignore this atom, seek to the end of it.
datastream.seek(atom.position + atom.size)
def process(infilename, limit=float('inf')):
"""
Convert a Quicktime/MP4 file for streaming by moving the metadata to
the front of the file. This method writes a new file.
If limit is set to something other than zero it will be used as the
number of bytes to write of the atoms following the moov atom. This
is very useful to create a small sample of a file with full headers,
which can then be used in bug reports and such.
"""
if isinstance(infilename, str):
datastream = open(infilename, "rb")
elif isinstance(infilename, bytes):
datastream = io.BytesIO(infilename)
else:
raise TypeError("infilename must be a filename, bytes or file-like object")
# Get the top level atom index
index = get_index(datastream)
mdat_pos = 999999
free_size = 0
# Make sure moov occurs AFTER mdat, otherwise no need to run!
for atom in index:
# The atoms are guaranteed to exist from get_index above!
if atom.name == "moov":
moov_atom = atom
moov_pos = atom.position
elif atom.name == "mdat":
mdat_pos = atom.position
elif atom.name == "free" and atom.position < mdat_pos:
# This free atom is before the mdat!
free_size += atom.size
log.info("Removing free atom at %d (%d bytes)" % (atom.position, atom.size))
elif atom.name == "\x00\x00\x00\x00" and atom.position < mdat_pos:
# This is some strange zero atom with incorrect size
free_size += 8
log.info("Removing strange zero atom at %s (8 bytes)" % atom.position)
# Offset to shift positions
offset = moov_atom.size - free_size
if moov_pos < mdat_pos:
# moov appears to be in the proper place, don't shift by moov size
offset -= moov_atom.size
if not free_size:
# No free atoms and moov is correct, we are done!
log.error("This file appears to already be setup for streaming!")
# Stupid hack to retrun the non-processed file:
if isinstance(infilename, str):
return open(infilename, "rb").read()
elif isinstance(infilename, bytes):
return io.BytesIO(infilename).read()
# Read and fix moov
moov = _patch_moov(datastream, moov_atom, offset)
log.info("Writing output...")
outfile = b''
# Write ftype
for atom in index:
if atom.name == "ftyp":
log.debug("Writing ftyp... (%d bytes)" % atom.size)
datastream.seek(atom.position)
outfile += datastream.read(atom.size)
# Write moov
_bytes = moov.getvalue()
log.debug("Writing moov... (%d bytes)" % len(_bytes))
outfile += _bytes
# Write the rest
atoms = [item for item in index if item.name not in ["ftyp", "moov", "free"]]
for atom in atoms:
log.debug("Writing %s... (%d bytes)" % (atom.name, atom.size))
datastream.seek(atom.position)
# for compatability, allow '0' to mean no limit
cur_limit = limit or float('inf')
cur_limit = min(cur_limit, atom.size)
for chunk in get_chunks(datastream, CHUNK_SIZE, cur_limit):
outfile += chunk
return outfile
def _patch_moov(datastream, atom, offset):
datastream.seek(atom.position)
moov = io.BytesIO(datastream.read(atom.size))
# reload the atom from the fixed stream
atom = _read_atom_ex(moov)
for atom in _find_atoms_ex(atom, moov):
# Read either 32-bit or 64-bit offsets
ctype, csize = dict(
stco=('L', 4),
co64=('Q', 8),
)[atom.name]
# Get number of entries
version, entry_count = struct.unpack(">2L", moov.read(8))
log.info("Patching %s with %d entries" % (atom.name, entry_count))
entries_pos = moov.tell()
struct_fmt = ">%(entry_count)s%(ctype)s" % vars()
# Read entries
entries = struct.unpack(struct_fmt, moov.read(csize * entry_count))
# Patch and write entries
offset_entries = [entry + offset for entry in entries]
moov.seek(entries_pos)
moov.write(struct.pack(struct_fmt, *offset_entries))
return moov
def get_chunks(stream, chunk_size, limit):
remaining = limit
while remaining:
chunk = stream.read(min(remaining, chunk_size))
if not chunk:
return
remaining -= len(chunk)
yield chunk
scraibe/audio.py
+150
View File
@@ -0,0 +1,150 @@
"""
Audio Processor Module
=======================
This module provides the AudioProcessor class, utilizing PyTorchaudio for handling audio files.
It includes functionalities to load, cut, and manage audio waveforms, offering efficient and
flexible audio processing.
Available Classes:
- AudioProcessor: Processes audio waveforms and provides methods for loading,
cutting, and handling audio.
Usage:
from .audio_import AudioProcessor
processor = AudioProcessor.from_file("path/to/audiofile.wav")
cut_waveform = processor.cut(start=1.0, end=5.0)
Constants:
- SAMPLE_RATE (int): Default sample rate for processing.
- NORMALIZATION_FACTOR (float): Normalization factor for audio waveform.
"""
from subprocess import CalledProcessError, run
import numpy as np
import torch
SAMPLE_RATE = 16000
NORMALIZATION_FACTOR = 32768.0
class AudioProcessor:
"""
Audio Processor class that leverages PyTorchaudio to provide functionalities
for loading, cutting, and handling audio waveforms.
Attributes:
waveform: torch.Tensor
The audio waveform tensor.
sr: int
The sample rate of the audio.
"""
def __init__(self, waveform: torch.Tensor, sr : int = SAMPLE_RATE,
*args, **kwargs) -> None:
"""
Initialize the AudioProcessor object.
Args:
waveform (torch.Tensor): The audio waveform tensor.
sr (int, optional): The sample rate of the audio. Defaults to SAMPLE_RATE.
args: Additional arguments.
kwargs: Additional keyword arguments, e.g., device to use for processing.
If CUDA is available, it defaults to CUDA.
Raises:
ValueError: If the provided sample rate is not of type int.
"""
device = kwargs.get("device", "cuda" if torch.cuda.is_available() else "cpu")
self.waveform = waveform.to(device)
self.sr = sr
if not isinstance(self.sr, int):
raise ValueError("Sample rate should be a single value of type int," \
f"not {len(self.sr)} and type {type(self.sr)}")
@classmethod
def from_file(cls, file: str, *args, **kwargs) -> 'AudioProcessor':
"""
Create an AudioProcessor instance from an audio file.
Args:
file (str): The audio file path.
Returns:
AudioProcessor: An instance of the AudioProcessor class containing the loaded audio.
"""
audio, sr = cls.load_audio(file , *args, **kwargs)
audio = torch.from_numpy(audio)
return cls(audio, sr)
def cut(self, start: float, end: float) -> torch.Tensor:
"""
Cut a segment from the audio waveform between the specified start and end times.
Args:
start (float): Start time in seconds.
end (float): End time in seconds.
Returns:
torch.Tensor: The cut waveform segment.
"""
start = int(start * self.sr)
if (isinstance(end, float) or isinstance(end, int)) and isinstance(self.sr, int):
end = int(np.ceil(end * self.sr))
else:
end = int(torch.ceil(end * self.sr))
return self.waveform[start:end]
@staticmethod
def load_audio(file: str, sr: int = SAMPLE_RATE):
"""
Open an audio file and read it as a mono waveform, resampling if necessary.
This method ensures compatibility with pyannote.audio
and requires the ffmpeg CLI in PATH.
Args:
file (str): The audio file to open.
sr (int, optional): The desired sample rate. Defaults to SAMPLE_RATE.
Returns:
tuple: A NumPy array containing the audio waveform in float32 dtype
and the sample rate.
Raises:
RuntimeError: If failed to load audio.
"""
# This launches a subprocess to decode audio while down-mixing
# and resampling as necessary. Requires the ffmpeg CLI in PATH.
# fmt: off
cmd = [
"ffmpeg",
"-nostdin",
"-threads", "0",
"-i", file,
"-f", "s16le",
"-ac", "1",
"-acodec", "pcm_s16le",
"-ar", str(sr),
"-"
]
# fmt: on
try:
out = run(cmd, capture_output=True, check=True).stdout
except CalledProcessError as e:
raise RuntimeError(f"Failed to load audio: {e.stderr.decode()}") from e
out = np.frombuffer(out, np.int16).flatten().astype(np.float32) / NORMALIZATION_FACTOR
return out , sr
def __repr__(self) -> str:
return f'TorchAudioProcessor(waveform={len(self.waveform)}, sr={int(self.sr)})'
scraibe/autotranscript.py
+290
View File
@@ -0,0 +1,290 @@
"""
Scraibe Class
--------------------
This class serves as the core of the transcription system, responsible for handling
transcription and diarization of audio files. It leverages pretrained models for
speech-to-text (such as Whisper) and speaker diarization (such as pyannote.audio),
providing an accessible interface for audio processing tasks such as transcription,
speaker separation, and timestamping.
By encapsulating the complexities of underlying models, it allows for straightforward
integration into various applications, ranging from transcription services to voice assistants.
Available Classes:
- Scraibe: Main class for performing transcription and diarization.
Includes methods for loading models, processing audio files,
and formatting the transcription output.
Usage:
from scraibe import Scraibe
model = Scraibe()
transcript = model.autotranscribe("path/to/audiofile.wav")
"""
# Standard Library Imports
import os
from glob import iglob
from subprocess import run
from typing import TypeVar, Union
from warnings import warn
# Third-Party Imports
import torch
from numpy import ndarray
from tqdm import trange
# Application-Specific Imports
from .audio import AudioProcessor
from .diarisation import Diariser
from .transcriber import Transcriber, whisper
from .transcript_exporter import Transcript
DiarisationType = TypeVar('DiarisationType')
class Scraibe:
"""
Scraibe is a class responsible for managing the transcription and diarization of audio files.
It serves as the core of the transcription system, incorporating pretrained models
for speech-to-text (such as Whisper) and speaker diarization (such as pyannote.audio),
allowing for comprehensive audio processing.
Attributes:
transcriber (Transcriber): The transcriber object to handle transcription.
diariser (Diariser): The diariser object to handle diarization.
Methods:
__init__: Initializes the Scraibe class with appropriate models.
transcribe: Transcribes an audio file using the whisper model and pyannote diarization model.
remove_audio_file: Removes the original audio file to avoid disk space issues or ensure data privacy.
get_audio_file: Gets an audio file as an AudioProcessor object.
"""
def __init__(self,
whisper_model: Union[bool, str, whisper] = None,
dia_model : Union[bool, str, DiarisationType] = None,
**kwargs) -> None:
"""Initializes the Scraibe class.
Args:
whisper_model (Union[bool, str, whisper], optional):
Path to whisper model or whisper model itself.
diarisation_model (Union[bool, str, DiarisationType], optional):
Path to pyannote diarization model or model itself.
**kwargs: Additional keyword arguments for whisper
and pyannote diarization models.
"""
if whisper_model is None:
self.transcriber = Transcriber.load_model("medium", **kwargs)
elif isinstance(whisper_model, str):
self.transcriber = Transcriber.load_model(whisper_model, **kwargs)
else:
self.transcriber = whisper_model
if dia_model is None:
self.diariser = Diariser.load_model(**kwargs)
elif isinstance(dia_model, str):
self.diariser = Diariser.load_model(dia_model, **kwargs)
else:
self.diariser = dia_model
if kwargs.get("verbose"):
print("Scraibe initialized all models successfully loaded.")
self.verbose = True
else:
self.verbose = False
def autotranscribe(self, audio_file : Union[str, torch.Tensor, ndarray],
remove_original : bool = False,
**kwargs) -> Transcript:
"""
Transcribes an audio file using the whisper model and pyannote diarization model.
Args:
audio_file (Union[str, torch.Tensor, ndarray]):
Path to audio file or a tensor representing the audio.
remove_original (bool, optional): If True, the original audio file will
be removed after transcription.
*args: Additional positional arguments for diarization and transcription.
**kwargs: Additional keyword arguments for diarization and transcription.
Returns:
Transcript: A Transcript object containing the transcription,
which can be exported to different formats.
"""
if kwargs.get("verbose"):
self.verbose = kwargs.get("verbose")
# Get audio file as an AudioProcessor object
audio_file = self.get_audio_file(audio_file)
# Prepare waveform and sample rate for diarization
dia_audio = {
"waveform" : audio_file.waveform.reshape(1,len(audio_file.waveform)),
"sample_rate": audio_file.sr
}
if self.verbose:
print("Starting diarisation.")
diarisation = self.diariser.diarization(dia_audio, **kwargs)
if not diarisation["segments"]:
print("No segments found. Try to run transcription without diarisation.")
transcript = self.transcriber.transcribe(audio_file.waveform, **kwargs)
final_transcript= {0 : {"speakers" : 'SPEAKER_01',
"segments" : [0, len(audio_file.waveform)],
"text" : transcript}}
return Transcript(final_transcript)
if self.verbose:
print("Diarisation finished. Starting transcription.")
audio_file.sr = torch.Tensor([audio_file.sr]).to(audio_file.waveform.device)
# Transcribe each segment and store the results
final_transcript = dict()
for i in trange(len(diarisation["segments"]), desc= "Transcribing", disable = not self.verbose):
seg = diarisation["segments"][i]
audio = audio_file.cut(seg[0], seg[1])
transcript = self.transcriber.transcribe(audio, **kwargs)
final_transcript[i] = {"speakers" : diarisation["speakers"][i],
"segments" : seg,
"text" : transcript}
# Remove original file if needed
if remove_original:
if kwargs.get("shred") is True:
self.remove_audio_file(audio_file, shred=True)
else:
self.remove_audio_file(audio_file, shred=False)
return Transcript(final_transcript)
def diarization(self, audio_file : Union[str, torch.Tensor, ndarray],
**kwargs) -> dict:
"""
Perform diarization on an audio file using the pyannote diarization model.
Args:
audio_file (Union[str, torch.Tensor, ndarray]):
The audio source which can either be a path to the audio file or a tensor representation.
**kwargs:
Additional keyword arguments for diarization.
Returns:
dict:
A dictionary containing the results of the diarization process.
"""
# Get audio file as an AudioProcessor object
audio_file = self.get_audio_file(audio_file)
# Prepare waveform and sample rate for diarization
dia_audio = {
"waveform" : audio_file.waveform.reshape(1,len(audio_file.waveform)),
"sample_rate": audio_file.sr
}
print("Starting diarisation.")
diarisation = self.diariser.diarization(dia_audio, **kwargs)
return diarisation
def transcribe(self, audio_file : Union[str, torch.Tensor, ndarray],
**kwargs):
"""
Transcribe the provided audio file.
Args:
audio_file (Union[str, torch.Tensor, ndarray]):
The audio source, which can either be a path or a tensor representation.
**kwargs:
Additional keyword arguments for transcription.
Returns:
str:
The transcribed text from the audio source.
"""
audio_file = self.get_audio_file(audio_file)
return self.transcriber.transcribe(audio_file.waveform, **kwargs)
@staticmethod
def remove_audio_file(audio_file : str,
shred : bool = False) -> None:
"""
Removes the original audio file to avoid disk space issues or ensure data privacy.
Args:
audio_file_path (str): Path to the audio file.
shred (bool, optional): If True, the audio file will be shredded,
not just removed.
"""
if not os.path.exists(audio_file):
raise ValueError(f"Audiofile {audio_file} does not exist.")
if shred:
warn("Shredding audiofile can take a long time.", RuntimeWarning)
gen = iglob(f'{audio_file}', recursive=True)
cmd = ['shred', '-zvu', '-n', '10', f'{audio_file}']
if os.path.isdir(audio_file):
raise ValueError(f"Audiofile {audio_file} is a directory.")
for file in gen:
print(f'shredding {file} now\n')
run(cmd , check=True)
else:
os.remove(audio_file)
print(f"Audiofile {audio_file} removed.")
@staticmethod
def get_audio_file(audio_file : Union[str, torch.Tensor, ndarray],
*args, **kwargs) -> AudioProcessor:
"""Gets an audio file as TorchAudioProcessor.
Args:
audio_file (Union[str, torch.Tensor, ndarray]): Path to the audio file or
a tensor representing the audio.
*args: Additional positional arguments.
**kwargs: Additional keyword arguments.
Returns:
AudioProcessor: An object containing the waveform and sample rate in
torch.Tensor format.
"""
if isinstance(audio_file, str):
audio_file = AudioProcessor.from_file(audio_file)
elif isinstance(audio_file, torch.Tensor):
audio_file = AudioProcessor(audio_file[0], audio_file[1])
elif isinstance(audio_file, ndarray):
audio_file = AudioProcessor(torch.Tensor(audio_file[0]),
audio_file[1])
if not isinstance(audio_file, AudioProcessor):
raise ValueError(f'Audiofile must be of type AudioProcessor,' \
f'not {type(audio_file)}')
return audio_file
def __repr__(self):
return f"Scraibe(transcriber={self.transcriber}, diariser={self.diariser})"
scraibe/cli.py
+167
View File
@@ -0,0 +1,167 @@
"""
Command-Line Interface (CLI) for the Scraibe class,
allowing for user interaction to transcribe and diarize audio files.
The function includes arguments for specifying the audio files, model paths,
output formats, and other options necessary for transcription.
"""
import os
from argparse import ArgumentParser, ArgumentDefaultsHelpFormatter
import json
from .autotranscript import Scraibe
from .app.gradio_app import gradio_Interface
from whisper.tokenizer import LANGUAGES , TO_LANGUAGE_CODE
from torch.cuda import is_available
from torch import set_num_threads
def cli():
"""
Command-Line Interface (CLI) for the Scraibe class, allowing for user interaction to transcribe
and diarize audio files. The function includes arguments for specifying the audio files, model paths,
output formats, and other options necessary for transcription.
This function can be executed from the command line to perform transcription tasks, providing a
user-friendly way to access the Scraibe class functionalities.
"""
def str2bool(string):
str2val = {"True": True, "False": False}
if string in str2val:
return str2val[string]
else:
raise ValueError(f"Expected one of {set(str2val.keys())}, got {string}")
parser = ArgumentParser(formatter_class = ArgumentDefaultsHelpFormatter)
group = parser.add_mutually_exclusive_group()
parser.add_argument("-f","--audio-files", nargs="+", type=str, default=None,
help="List of audio files to transcribe.")
group.add_argument('--start-server', action='store_true',
help='Start the Gradio app.')
parser.add_argument("--port", type=int, default= None,
help="Port to run the Gradio app on. Defaults to 7860.")
parser.add_argument("--server-name", type=str, default= None,
help="Name of the Gradio app. If empty 127.0.0.1 or 0.0.0.0 will be used.")
parser.add_argument("--whisper-model-name", default="medium",
help="Name of the Whisper model to use.")
parser.add_argument("--whisper-model-directory", type=str, default= None,
help="Path to save Whisper model files; defaults to ./models/whisper.")
parser.add_argument("--diarization-directory", type=str, default= None,
help="Path to the diarization model directory.")
parser.add_argument("--hf-token", default= None, type=str,
help="HuggingFace token for private model download.")
parser.add_argument("--inference-device",
default="cuda" if is_available() else "cpu",
help="Device to use for PyTorch inference.")
parser.add_argument("--num-threads", type=int, default=0,
help="Number of threads used by torch for CPU inference; overrides MKL_NUM_THREADS/OMP_NUM_THREADS.")
parser.add_argument("--output-directory", "-o", type=str, default=".",
help="Directory to save the transcription outputs.")
parser.add_argument("--output-format", "-of", type=str, default="txt",
choices=["txt", "json", "md", "html"],
help="Format of the output file; defaults to txt.")
parser.add_argument("--verbose-output", type=str2bool, default=True,
help="Enable or disable progress and debug messages.")
parser.add_argument("--task", type=str, default= 'autotranscribe', # unifinished code
choices=["autotranscribe", "diarization",
"autotranscribe+translate", "translate", 'transcribe'],
help="Choose to perform transcription, diarization, or translation. \
If set to translate, the output will be translated to English.")
parser.add_argument("--language", type=str, default=None,
choices=sorted(LANGUAGES.keys()) + sorted([k.title() for k in TO_LANGUAGE_CODE.keys()]),
help="Language spoken in the audio. Specify None to perform language detection.")
args = parser.parse_args()
arg_dict = vars(args)
# configure output
out_folder = arg_dict.pop("output_directory")
os.makedirs(out_folder, exist_ok=True)
out_format = arg_dict.pop("output_format")
# seup server arg:
start_server = arg_dict.pop("start_server")
task = arg_dict.pop("task")
if args.num_threads > 0:
set_num_threads(arg_dict.pop("num_threads"))
class_kwargs = {'whisper_model' : arg_dict.pop("whisper_model_name"),
'dia_model': arg_dict.pop("diarization_directory"),
'use_auth_token' : arg_dict.pop("hf_token")}
if arg_dict["whisper_model_directory"]:
class_kwargs["download_root"] = arg_dict.pop("whisper_model_directory")
model = Scraibe(**class_kwargs)
if arg_dict["audio_files"]:
audio_files = arg_dict.pop("audio_files")
if task == "autotranscribe" or task == "autotranscribe+translate":
for audio in audio_files:
if task == "autotranscribe+translate":
task = "translate"
else:
task = "transcribe"
out = model.autotranscribe(audio,task = task, language=arg_dict.pop("language"), verbose = arg_dict.pop("verbose_output"))
basename = audio.split("/")[-1].split(".")[0]
print(f'Saving {basename}.{out_format} to {out_folder}')
out.save(os.path.join(out_folder, f"{basename}.{out_format}"))
elif task == "diarization":
for audio in audio_files:
if arg_dict.pop("verbose_output"):
print(f"Verbose not implemented for diarization.")
out = model.diarization(audio)
basename = audio.split("/")[-1].split(".")[0]
path = os.path.join(out_folder, f"{basename}.{out_format}")
print(f'Saving {basename}.{out_format} to {out_folder}')
with open(path, "w") as f:
json.dump(json.dumps(out, indent= 1), f)
elif task == "transcribe" or task == "translate":
for audio in audio_files:
out = model.transcribe(audio, task = task,
language= arg_dict.pop("language"),
verbose = arg_dict.pop("verbose_output"))
basename = audio.split("/")[-1].split(".")[0]
path = os.path.join(out_folder, f"{basename}.{out_format}")
with open(path, "w") as f:
f.write(out)
if start_server: # unfinished code
gradio_Interface(model).queue().launch(server_port=args.port, server_name=args.server_name)
if __name__ == "__main__":
cli()
scraibe/diarisation.py
+247
View File
@@ -0,0 +1,247 @@
"""
Diarisation Class
------------------
This class serves as the heart of the speaker diarization system, responsible for identifying
and segmenting individual speakers from a given audio file. It leverages a pretrained model
from pyannote.audio, providing an accessible interface for audio processing tasks such as
speaker separation, and timestamping.
By encapsulating the complexities of the underlying model, it allows for straightforward
integration into various applications, ranging from transcription services to voice assistants.
Available Classes:
- Diariser: Main class for performing speaker diarization.
Includes methods for loading models, processing audio files,
and formatting the diarization output.
Constants:
- TOKEN_PATH (str): Path to the Pyannote token.
- PYANNOTE_DEFAULT_PATH (str): Default path to Pyannote models.
- PYANNOTE_DEFAULT_CONFIG (str): Default configuration for Pyannote models.
Usage:
from .diarisation import Diariser
model = Diariser.load_model(model="path/to/model/config.yaml")
diarisation_output = model.diarization("path/to/audiofile.wav")
"""
import os
from pathlib import Path
from typing import TypeVar, Union
from pyannote.audio import Pipeline
from pyannote.audio.pipelines.speaker_diarization import SpeakerDiarization
from torch import Tensor
from .misc import PYANNOTE_DEFAULT_PATH, PYANNOTE_DEFAULT_CONFIG
Annotation = TypeVar('Annotation')
TOKEN_PATH = os.path.join(os.path.dirname(
os.path.realpath(__file__)), '.pyannotetoken')
class Diariser:
"""
Handles the diarization process of an audio file using a pretrained model
from pyannote.audio. Diarization is the task of determining "who spoke when."
Args:
model: The pretrained model to use for diarization.
"""
def __init__(self, model) -> None:
self.model = model
def diarization(self, audiofile : Union[str, Tensor, dict] ,
*args, **kwargs) -> Annotation:
"""
Perform speaker diarization on the provided audio file,
effectively separating different speakers
and providing a timestamp for each segment.
Args:
audiofile: The path to the audio file or a torch.Tensor
containing the audio data.
args: Additional arguments for the diarization model.
kwargs: Additional keyword arguments for the diarization model.
Returns:
dict: A dictionary containing speaker names,
segments, and other information related
to the diarization process.
"""
kwargs = self._get_diarisation_kwargs(**kwargs)
diarization = self.model(audiofile,*args, **kwargs)
out = self.format_diarization_output(diarization)
return out
@staticmethod
def format_diarization_output(dia : Annotation) -> dict:
"""
Formats the raw diarization output into a more usable structure for this project.
Args:
dia: Raw diarization output.
Returns:
dict: A structured representation of the diarization, with speaker names
as keys and a list of tuples representing segments as values.
"""
dia_list = list(dia.itertracks(yield_label=True))
diarization_output = {"speakers": [], "segments": []}
normalized_output = []
index_start_speaker = 0
index_end_speaker = 0
current_speaker = str()
###
# Sometimes two consecutive speakers are the same
# This loop removes these duplicates
###
if len(dia_list) == 1:
normalized_output.append([0, 0, dia_list[0][2]])
else:
for i, (_, _, speaker) in enumerate(dia_list):
if i == 0:
current_speaker = speaker
if speaker != current_speaker:
index_end_speaker = i - 1
normalized_output.append([index_start_speaker,
index_end_speaker,
current_speaker])
index_start_speaker = i
current_speaker = speaker
if i == len(dia_list) - 1:
index_end_speaker = i
normalized_output.append([index_start_speaker,
index_end_speaker,
current_speaker])
for outp in normalized_output:
start = dia_list[outp[0]][0].start
end = dia_list[outp[1]][0].end
diarization_output["segments"].append([start, end])
diarization_output["speakers"].append(outp[2])
return diarization_output
@staticmethod
def _get_token():
"""
Retrieves the Huggingface token from a local file. This token is required
for accessing certain online resources.
Raises:
ValueError: If the token is not found.
Returns:
str: The Huggingface token.
"""
if os.path.exists(TOKEN_PATH):
with open(TOKEN_PATH, 'r', encoding="utf-8") as file:
token = file.read()
else:
raise ValueError('No token found.' \
'Please create a token at https://huggingface.co/settings/token' \
f'and save it in a file called {TOKEN_PATH}')
return token
@staticmethod
def _save_token(token):
"""
Saves the provided Huggingface token to a local file. This facilitates future
access to online resources without needing to repeatedly authenticate.
Args:
token: The Huggingface token to save.
"""
with open(TOKEN_PATH, 'w', encoding="utf-8") as file:
file.write(token)
@classmethod
def load_model(cls,
model: str = PYANNOTE_DEFAULT_CONFIG,
use_auth_token: str = None,
cache_token: bool = True,
cache_dir: Union[Path, str] = PYANNOTE_DEFAULT_PATH,
hparams_file: Union[str, Path] = None,
*args, **kwargs
) -> Pipeline:
"""
Loads a pretrained model from pyannote.audio,
either from a local cache or online repository.
Args:
model: Path or identifier for the pyannote model.
default: /models/pyannote/speaker_diarization/config.yaml
token: Optional HUGGINGFACE_TOKEN for authenticated access.
cache_token: Whether to cache the token locally for future use.
cache_dir: Directory for caching models.
hparams_file: Path to a YAML file containing hyperparameters.
args: Additional arguments only to avoid errors.
kwargs: Additional keyword arguments only to avoid errors.
Returns:
Pipeline: A pyannote.audio Pipeline object, encapsulating the loaded model.
"""
if cache_token and use_auth_token is not None:
cls._save_token(use_auth_token)
if not os.path.exists(model) and use_auth_token is None:
use_auth_token = cls._get_token()
model = 'pyannote/speaker-diarization'
elif not os.path.exists(model) and use_auth_token is not None:
model = 'pyannote/speaker-diarization'
_model = Pipeline.from_pretrained(model,
use_auth_token = use_auth_token,
cache_dir = cache_dir,
hparams_file = hparams_file,)
if _model is None:
raise ValueError('Unable to load model either from local cache' \
'or from huggingface.co models. Please check your token' \
'or your local model path')
return cls(_model)
@staticmethod
def _get_diarisation_kwargs(**kwargs) -> dict:
"""
Validates and extracts the keyword arguments for the pyannote diarization model.
Ensures that the provided keyword arguments match the expected parameters,
filtering out any invalid or unnecessary arguments.
Returns:
dict: A dictionary containing the validated keyword arguments.
"""
_possible_kwargs = SpeakerDiarization.apply.__code__.co_varnames
diarisation_kwargs = {k: v for k, v in kwargs.items() if k in _possible_kwargs}
return diarisation_kwargs
def __repr__(self):
return f"Diarisation(model={self.model})"
scraibe/misc.py
+40
View File
@@ -0,0 +1,40 @@
import os
import yaml
from pyannote.audio.core.model import CACHE_DIR as PYANNOTE_CACHE_DIR
CACHE_DIR = os.getenv(
"AUTOT_CACHE",
os.path.expanduser("~/.cache/torch/models"),
)
if CACHE_DIR != PYANNOTE_CACHE_DIR:
os.environ["PYANNOTE_CACHE"] = os.path.join(CACHE_DIR, "pyannote")
WHISPER_DEFAULT_PATH = os.path.join(CACHE_DIR, "whisper")
PYANNOTE_DEFAULT_PATH = os.path.join(CACHE_DIR, "pyannote")
PYANNOTE_DEFAULT_CONFIG = os.path.join(PYANNOTE_DEFAULT_PATH, "config.yaml")
def config_diarization_yaml(file_path: str, path_to_segmentation: str = None) -> None:
"""Configure diarization pipeline from a YAML file.
This function updates the YAML file to use the given segmentation model
offline, and avoids manual file manipulation.
Args:
file_path (str): Path to the YAML file.
path_to_segmentation (str, optional): Optional path to the segmentation model.
Raises:
FileNotFoundError: If the segmentation model file is not found.
"""
with open(file_path, "r") as stream:
yml = yaml.safe_load(stream)
segmentation_path = path_to_segmentation or os.path.join(PYANNOTE_DEFAULT_PATH, "pytorch_model.bin")
yml["pipeline"]["params"]["segmentation"] = segmentation_path
if not os.path.exists(segmentation_path):
raise FileNotFoundError(f"Segmentation model not found at {segmentation_path}")
with open(file_path, "w") as stream:
yaml.dump(yml, stream)
scraibe/transcriber.py
+182
View File
@@ -0,0 +1,182 @@
"""
Transcriber Module
------------------
This module provides the Transcriber class, a comprehensive tool for working with Whisper models.
The Transcriber class offers functionalities such as loading different Whisper models, transcribing audio files,
and saving transcriptions to text files. It acts as an interface between various Whisper models and the user,
simplifying the process of audio transcription.
Main Features:
- Loading different sizes and versions of Whisper models.
- Transcribing audio in various formats including str, Tensor, and nparray.
- Saving the transcriptions to the specified paths.
- Adaptable to various language specifications.
- Options to control the verbosity of the transcription process.
Constants:
WHISPER_DEFAULT_PATH: Default path for downloading and loading Whisper models.
Usage:
>>> from your_package import Transcriber
>>> transcriber = Transcriber.load_model(model="medium")
>>> transcript = transcriber.transcribe(audio="path/to/audio.wav")
>>> transcriber.save_transcript(transcript, "path/to/save.txt")
"""
from whisper import Whisper, load_model
from typing import TypeVar , Union , Optional
from torch import Tensor, device
from numpy import ndarray
from .misc import WHISPER_DEFAULT_PATH
whisper = TypeVar('whisper')
class Transcriber:
"""
Transcriber Class
-----------------
The Transcriber class serves as a wrapper around Whisper models for efficient audio
transcription. By encapsulating the intricacies of loading models, processing audio,
and saving transcripts, it offers an easy-to-use interface
for users to transcribe audio files.
Attributes:
model (whisper): The Whisper model used for transcription.
Methods:
transcribe: Transcribes the given audio file.
save_transcript: Saves the transcript to a file.
load_model: Loads a specific Whisper model.
_get_whisper_kwargs: Private method to get valid keyword arguments for the whisper model.
Examples:
>>> transcriber = Transcriber.load_model(model="medium")
>>> transcript = transcriber.transcribe(audio="path/to/audio.wav")
>>> transcriber.save_transcript(transcript, "path/to/save.txt")
Note:
The class supports various sizes and versions of Whisper models. Please refer to
the load_model method for available options.
"""
def __init__(self, model: whisper ) -> None:
"""
Initialize the Transcriber class with a Whisper model.
Args:
model (whisper): The Whisper model to use for transcription.
"""
self.model = model
def transcribe(self, audio : Union[str, Tensor, ndarray] ,
*args, **kwargs) -> str:
"""
Transcribe an audio file.
Args:
audio (Union[str, Tensor, nparray]): The audio file to transcribe.
*args: Additional arguments.
**kwargs: Additional keyword arguments,
such as the language of the audio file.
Returns:
str: The transcript as a string.
"""
kwargs = self._get_whisper_kwargs(**kwargs)
if not kwargs.get("verbose"):
kwargs["verbose"] = None
result = self.model.transcribe(audio, *args, **kwargs)
return result["text"]
@staticmethod
def save_transcript(transcript : str , save_path : str) -> None:
"""
Save a transcript to a file.
Args:
transcript (str): The transcript as a string.
save_path (str): The path to save the transcript.
Returns:
None
"""
with open(save_path, 'w') as f:
f.write(transcript)
print(f'Transcript saved to {save_path}')
@classmethod
def load_model(cls,
model: str = "medium",
download_root: str = WHISPER_DEFAULT_PATH,
device: Optional[Union[str, device]] = None,
in_memory: bool = False,
*args, **kwargs
) -> 'Transcriber':
"""
Load whisper model.
Args:
model (str): Whisper model. Available models include:
- 'tiny.en'
- 'tiny'
- 'base.en'
- 'base'
- 'small.en'
- 'small'
- 'medium.en'
- 'medium'
- 'large-v1'
- 'large-v2'
- 'large'
download_root (str, optional): Path to download the model.
Defaults to WHISPER_DEFAULT_PATH.
device (Optional[Union[str, torch.device]], optional):
Device to load model on. Defaults to None.
in_memory (bool, optional): Whether to load model in memory.
Defaults to False.
args: Additional arguments only to avoid errors.
kwargs: Additional keyword arguments only to avoid errors.
Returns:
Transcriber: A Transcriber object initialized with the specified model.
"""
_model = load_model(model, download_root=download_root,
device=device, in_memory=in_memory)
return cls(_model)
@staticmethod
def _get_whisper_kwargs(**kwargs) -> dict:
"""
Get kwargs for whisper model. Ensure that kwargs are valid.
Returns:
dict: Keyword arguments for whisper model.
"""
_possible_kwargs = Whisper.transcribe.__code__.co_varnames
whisper_kwargs = {k: v for k, v in kwargs.items() if k in _possible_kwargs}
if (task := kwargs.get("task")):
whisper_kwargs["task"] = task
if (language := kwargs.get("language")):
whisper_kwargs["language"] = language
return whisper_kwargs
def __repr__(self) -> str:
return f"Transcriber(model={self.model})"
scraibe/transcript_exporter.py
+303
View File
@@ -0,0 +1,303 @@
import json
import time
from traceback import print_stack
from typing import Union
ALPHABET = [*"abcdefghijklmnopqrstuvwxyz"]
class Transcript:
"""
Class for storing transcript data, including speaker information and text segments,
and exporting it to various file formats such as JSON, HTML, and LaTeX.
"""
def __init__(self, transcript: dict) -> None:
"""
Initializes the Transcript object with the given transcript data.
Args:
transcript (dict): A dictionary containing the formatted transcript string.
Keys should correspond to segment IDs, and values should
contain speaker and segment information.
"""
self.transcript = transcript
self.speakers = self._extract_speakers()
self.segments = self._extract_segments()
self.annotation = {}
def annotate(self, *args, **kwargs) -> dict:
"""
Annotates the transcript to associate specific names with speakers.
Args:
args (list): List of speaker names. These will be mapped sequentially to the speakers.
kwargs (dict): Dictionary with speaker names as keys and list of segments as values.
Returns:
dict: Dictionary with speaker names as keys and list of segments as values.
Raises:
ValueError: If the number of speaker names does not match the number
of speakers, or if an unknown speaker is found.
"""
annotations = {}
if args and len(args) != len(self.speakers):
raise ValueError("Number of speaker names does not match number of speakers")
if args:
for arg, speaker in zip(args, sorted(self.speakers)):
annotations[speaker] = arg
invalid_speakers = set(kwargs.keys()) - set(self.speakers)
if invalid_speakers:
raise ValueError(f"These keys are not speakers: {', '.join(invalid_speakers)}")
annotations.update({key: kwargs[key] for key in self.speakers if key in kwargs})
self.annotation = annotations
return self
def _extract_speakers(self) -> list:
"""
Extracts the unique speaker names from the transcript.
Returns:
list: List of unique speaker names in the transcript.
"""
return list(set([self.transcript[id]["speakers"] for id in self.transcript]))
def _extract_segments(self) -> list:
"""
Extracts all the text segments from the transcript.
Returns:
list: List of segments, where each segment is represented
by the starting and ending times.
"""
return [self.transcript[id]["segments"] for id in self.transcript]
def __str__(self) -> str:
"""
Converts the transcript to a string representation.
Returns:
str: String representation of the transcript, including speaker names and
time stamps for each segment.
"""
fstring = ""
for _id in self.transcript:
seq = self.transcript[_id]
if self.annotation:
speaker = self.annotation[seq["speakers"]]
else:
speaker = seq["speakers"]
segm = seq["segments"]
sseg = time.strftime("%H:%M:%S",time.gmtime(segm[0]))
eseg = time.strftime("%H:%M:%S",time.gmtime(segm[1]))
fstring += f"{speaker} ({sseg} ; {eseg}):\t{seq['text']}\n"
return fstring
def __repr__(self) -> str:
"""Return a string representation of the Transcript object.
Returns:
str: A string that provides an informative description of the object.
"""
return f"Transcript(speakers = {self.speakers},"\
f"segments = {self.segments}, annotation = {self.annotation})"
def get_dict(self) -> dict:
"""
Get transcript as dict
:return: transcript as dict
:rtype: dict
"""
return self.transcript
def get_json(self, *args, use_annotation : bool = True, **kwargs) -> str:
"""
Get transcript as json string
:return: transcript as json string
:rtype: str
"""
if "indent" not in kwargs:
kwargs["indent"] = 3
if use_annotation and self.annotation:
for _id in self.transcript:
seq = self.transcript[_id]
seq["speakers"] = self.annotation[seq["speakers"]]
return json.dumps(self.transcript, *args, **kwargs)
def get_html(self) -> str:
"""
Get transcript as html string
:return: transcript as html string
:rtype: str
"""
html = "<p>" + self.__str__().replace("\n", "<br>") + "</p>"
html = "<html><body>" + html + "</body></html>"
html = html.replace("\t", "&nbsp;&nbsp;&nbsp;&nbsp;")
return html
def get_md(self) -> str:
"""Get transcript as Markdown string, using HTML formatting.
Returns:
str: Transcript as a Markdown string.
"""
return self.get_html()
def get_tex(self) -> str:
"""Get transcript as LaTeX string. If no annotations are present, the speakers will
be annotated with the first letters of the alphabet.
Returns:
str: Transcript as LaTeX string.
"""
if not self.annotation:
self.annotate(*ALPHABET[:len(self.speakers)])
fstring ="\\begin{drama}"
for speaker in self.speakers:
fstring += "\n\t\\Character{"+ str(self.annotation[speaker]) + "}" \
"{"+ str(self.annotation[speaker]) + "}"
for id in self.transcript:
seq = self.transcript[id]
speaker = self.annotation[seq["speakers"]]
fstring += f"\n\\{speaker}speaks:\n{seq['text']}"
fstring += "\n\\end{drama}"
return fstring
def to_json(self,path, *args, **kwargs) -> None:
"""Save transcript as json file
Args:
path (str): path to save file
"""
with open(path, "w") as f:
json.dump(self.transcript, f, *args, **kwargs)
def to_txt(self, path: str) -> None:
"""Save transcript as a LaTeX file (placeholder function, implementation needed).
Args:
path (str): Path to save the LaTeX file.
"""
with open(path, "w") as f:
f.write(self.__str__())
def to_md(self, path: str) -> None:
"""Get transcript as Markdown string, using HTML formatting.
Returns:
str: Transcript as a Markdown string.
"""
return self.to_html(path)
def to_html(self, path: str) -> None:
"""
Save transcript as html file
:param path: path to save file
:type path: str
"""
with open(path, "w") as file:
file.write(self.get_html())
def to_tex(self, path: str) -> None:
"""Save transcript as a LaTeX file (placeholder function, implementation needed).
Args:
path (str): Path to save the LaTeX file.
"""
pass
def to_pdf(self, path: str) -> None:
"""Save transcript as a PDF file (placeholder function, implementation needed).
Args:
path (str): Path to save the PDF file.
"""
pass
def save(self, path: str, *args, **kwargs) -> None:
"""Save transcript to file with the given path and file format.
This method can save the transcript in various formats including JSON, TXT,
MD, HTML, TEX, and PDF. The file format is determined by the extension of
the path.
Args:
path (str): Path to save the file, including the desired file extension.
*args: Additional positional arguments to be passed to the specific save methods.
**kwargs: Additional keyword arguments to be passed to the specific save methods.
Raises:
ValueError: If the file format specified in the path is unknown.
"""
if path.endswith(".json"):
self.to_json(path, *args, **kwargs)
elif path.endswith(".txt"):
self.to_txt(path, *args, **kwargs)
elif path.endswith(".md"):
self.to_md(path, *args, **kwargs)
elif path.endswith(".html"):
self.to_html(path, *args, **kwargs)
elif path.endswith(".tex"):
self.to_tex(path, *args, **kwargs)
elif path.endswith(".pdf"):
self.to_pdf(path, *args, **kwargs)
else:
raise ValueError("Unknown file format")
@classmethod
def from_json(cls, json: Union[dict, str]) -> "Transcript":
"""Load transcript from json file
Args:
path (str): path to json file
Returns:
Transcript: Transcript object
"""
if isinstance(json, dict):
return cls(json)
else:
try:
transcript = json.loads(json)
except:
with open(json, "r") as f:
transcript = json.load(f)
return cls(transcript)
autotranscript/version.py → scraibe/version.py
+2 -2
View File
@@ -1,8 +1,8 @@
import os
import subprocess as sp
MAJOR = 1
MINOR = 0
MAJOR = 0
MINOR = 1
MICRO = 0
MICRO_POST = 0
ISRELEASED = False
setup.cfg
+31
View File
@@ -0,0 +1,31 @@
[metadata]
name = scraibe
version = attr: scraibe.__version__
author = Jacob Schmieder
author_email = Jacob.Schmieder@dbfz.de
description = My package description
long_description = file: README.md, LICENSE
platforms = Linux
keywords = transcription speech recognition whisper pyannote audio speech-to-text speech-to-text transcription speech-to-text recognition voice-to-speech
license = GPL-3.0
classifiers =
Development Status :: 3 - Alpha
Environment :: GPU :: NVIDIA CUDA :: 11.2
License :: OSI Approved :: Open Software License 3.0 (OSL-3.0)
Topic :: Scientific/Engineering :: Artificial Intelligence
Programming Language :: Python :: 3.8
Programming Language :: Python :: 3.9
Programming Language :: Python :: 3.10
[options]
zip_safe = False
include_package_data = True
packages = find:
python_requires = >=3.7
install_requires =
requests
importlib-metadata; python_version<"3.8"
[options.entry_points]
console_scripts =
executable-name = scraibe.cli:cli
setup.py
+42 -20
View File
@@ -1,9 +1,10 @@
from calendar import c
import pkg_resources
import os
from setuptools import setup, find_packages
module_name = "autotranscript"
github_url = "https://github.com/Jaikinator/transcriptor"
module_name = "scraibe"
github_url = "https://github.com/JSchmie/autotranscript"
file_dir = os.path.dirname(os.path.realpath(__file__))
absdir = lambda p: os.path.join(file_dir, p)
@@ -15,24 +16,45 @@ version = {"__file__": verfile}
with open(verfile, "r") as fp:
exec(fp.read(), version)
############### setup ###############
build_version = "OPTB_BUILD" in os.environ
build_version = "SCRAIBE_BUILD" in os.environ
setup(
name=module_name,
version=version["get_version"](build_version),
packages=find_packages(),
python_requires="~=3.9",
readme="README.md",
install_requires = [str(r) for r in pkg_resources.parse_requirements(
open(os.path.join(os.path.dirname(__file__), "requirements.txt"))
)
],
url= github_url,
license='',
author='Jacob Schmieder',
author_email='',
description='Transcription tool for audio files based on Whisper',
#entry_points={'console_scripts': ['autotranscript = autotranscript.__main__:main']}
)
if __name__ == "__main__":
setup(
name=module_name,
version=version["get_version"](build_version),
packages=find_packages(),
python_requires=">=3.8",
readme="README.md",
install_requires = [str(r) for r in pkg_resources.parse_requirements(
open(os.path.join(os.path.dirname(__file__), "requirements.txt"))
)
],
dependency_links=[
'https://download.pytorch.org/whl/cu113',
],
url= github_url,
license='GPL-3',
author='Jacob Schmieder',
author_email='Jacob.Schmieder@dbfz.de',
description='Transcription tool for audio files based on Whisper and Pyannote',
classifiers=[
'Development Status :: 3 - Alpha',
'Environment :: GPU :: NVIDIA CUDA :: 11.2',
'License :: OSI Approved :: Open Software License 3.0 (OSL-3.0)',
'Topic :: Scientific/Engineering :: Artificial Intelligence',
'Programming Language :: Python :: 3.8',
'Programming Language :: Python :: 3.9',
'Programming Language :: Python :: 3.10'],
keywords = ['transcription', 'speech recognition', 'whisper', 'pyannote', 'audio',
'speech-to-text', 'speech-to-text transcription', 'speech-to-text recognition',
'voice-to-speech'],
package_data={'scraibe.app' : ["*.html", "*.svg"]},
entry_points={'console_scripts':
['scraibe = scraibe.cli:cli']}
)
tests/test_autotranscript.py
+120
View File
@@ -0,0 +1,120 @@
import pytest
from scraibe import Transcriber
from unittest.mock import patch, mock_open
import os
def test_load_pyannote_model():
"""
Test load_pyannote_test
"""
from pyannote.audio.pipelines.speaker_diarization import SpeakerDiarization
from pyannote.audio import Pipeline
pipeline = Pipeline.from_pretrained("models/pyannote/speaker_diarization/config.yaml")
assert isinstance(pipeline, SpeakerDiarization)
# Test Transcribtion class
@pytest.fixture
def transcriber():
"""
Prepare Transcriber for testing
Returns: Transcriber Object
"""
return Transcriber.load_model("medium", local=True)
def test_Transcriber_init(transcriber):
"""
Test Transcriber initialization with a whisper model
"""
assert isinstance(transcriber, Transcriber)
def test_transcription(transcriber):
"""
Test transcription
"""
transcript = transcriber.transcribe("tests/test.wav")
assert isinstance(transcript, str)
def test_save_transcript_to_file(transcriber):
"""
Test save_transcript_to_file
"""
transcript = transcriber.transcribe("tests/test.wav")
Transcriber.save_transcript(transcript, "tests/output.txt")
assert os.path.exists("tests/output.txt")
os.remove("tests/output.txt")
# Test Diaraization class
from scraibe import Diariser
@pytest.fixture
def diarisation():
"""
Prepare Diarisation for testing
Returns: Diarisation Object
"""
return Diariser.load_model("models/pyannote/speaker_diarization/config.yaml", local=True)
def test_Diarisation_init(diarisation):
"""
Test Diarisation initialization with a pyannote model
"""
assert isinstance(diarisation, Diariser)
def test_diarisation(diarisation):
"""
Test diarisation
"""
diarisation = diarisation.diarization("tests/test.wav")
assert isinstance(diarisation, dict)
# Test AudioProcessor
from scraibe import AudioProcessor , TorchAudioProcessor
def test_AudioProcessor_init():
"""
Test AudioProcessor initialization
"""
audio = AudioProcessor("tests/test.wav")
assert isinstance(audio, AudioProcessor)
def test_AudioProcessor_convert():
"""
Test AudioProcessor convert
"""
audio = AudioProcessor("tests/test.wav")
audio.convert_audio("tests/test.mp3", format="mp3")
assert os.path.exists("tests/test.mp3")
def test_TorchAudioProcessor_from_file():
"""
Test TorchAudioProcessor initialization
"""
audio = TorchAudioProcessor.from_file("tests/test.wav")
assert isinstance(audio, TorchAudioProcessor)
os.remove("tests/test.mp3")
def test_TorchAudioProcessor_from_ffmpeg():
"""
Test TorchAudioProcessor initialization
"""
audio = TorchAudioProcessor.from_ffmpeg("tests/test.wav")
assert isinstance(audio, TorchAudioProcessor)
transcribe.py
-3
View File
@@ -1,3 +0,0 @@
from autotranscript import AutoTranscribe
AutoTranscribe(diarisation=True).transcribe()