Giter Site home page Giter Site logo

flakm / jupiter-search Goto Github PK

View Code? Open in Web Editor NEW
11.0 11.0 1.0 11.83 MB

Convert podstast RSS feed to transcriptions using whisper model

License: Apache License 2.0

Rust 92.26% Shell 2.01% Nix 3.32% Dockerfile 2.40%
podcast rust transcription

jupiter-search's People

Contributors

eraden avatar flakm avatar

Stargazers

 avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar

Watchers

 avatar  avatar  avatar

Forkers

eraden

jupiter-search's Issues

Deprecate ffmpeg subprocess calling and use fully rust solution

Currently in ffmpeg_decoder.rs conversion from mp3 to wav is done using creating a child subprocess and calling ffmpeg directly.
This is suboptimal for a number of reasons:

  1. It is prone to ffmpeg version changes (or it is missing completely)
  2. It requires the data to be written and read from the disk - downloaded mp3, converted/resampled waw
  3. It makes it impossible to use in no-std contexts like wasm

I've tried to prepare an alternative rust-only solution inside decoder.rs but it is not working - possibly because I'm a complete tool when it comes to audio.

Why it is a no-stopper:

  1. It's still a poc
  2. ffmpeg is pretty stable so the flags won't change from version to version
  3. The error will suggest that the ffmpeg is missing (?)

Prepare caching for transcription based on rss feed

Since transcription is very time-consuming (0.66xaudio_length) the results of stt should be cached based on RSS hashing to s3 or some other storage.

  • preferably cache should contain all the metadata for possible transformations (ie speaker tagging etc)
  • there should be some kind of index of all cached entries to enable watching new RSS feeds

RUSTSEC-2020-0071: Potential segfault in the time crate

Potential segfault in the time crate

Details
Package time
Version 0.1.44
URL time-rs/time#293
Date 2020-11-18
Patched versions >=0.2.23
Unaffected versions =0.2.0,=0.2.1,=0.2.2,=0.2.3,=0.2.4,=0.2.5,=0.2.6

Impact

Unix-like operating systems may segfault due to dereferencing a dangling pointer in specific circumstances. This requires an environment variable to be set in a different thread than the affected functions. This may occur without the user's knowledge, notably in a third-party library.

The affected functions from time 0.2.7 through 0.2.22 are:

  • time::UtcOffset::local_offset_at
  • time::UtcOffset::try_local_offset_at
  • time::UtcOffset::current_local_offset
  • time::UtcOffset::try_current_local_offset
  • time::OffsetDateTime::now_local
  • time::OffsetDateTime::try_now_local

The affected functions in time 0.1 (all versions) are:

  • at
  • at_utc
  • now

Non-Unix targets (including Windows and wasm) are unaffected.

Patches

Pending a proper fix, the internal method that determines the local offset has been modified to always return None on the affected operating systems. This has the effect of returning an Err on the try_* methods and UTC on the non-try_* methods.

Users and library authors with time in their dependency tree should perform cargo update, which will pull in the updated, unaffected code.

Users of time 0.1 do not have a patch and should upgrade to an unaffected version: time 0.2.23 or greater or the 0.3 series.

Workarounds

No workarounds are known.

See advisory page for additional details.

Investigate why stt unit test is hanging

Find out why unit test that does the same thing as the example code in get_transcript.rs hangs? oO

    fn stt_works() {
        let mut ctx = SttContext::try_new("resources/ggml-tiny.en.bin").unwrap();
        let t = ctx
            .get_transcript_file("resources/super_short.mp3", false)
            .unwrap();
        println!("{:?}", t);
        assert!(t.utterances.len() > 0)
    }

Prepare a new scorer for STT inference

Currently the results of generic models is less than satisfactory.
The language model clearly doesn't have any linuxy techy words, so it should be tuned for this specific purpose.

Training requires checkpoints that can be taken from the releases page https://github.com/coqui-ai/STT/releases?q=1.3.0&expanded=true all 1.x.x releases are backward compatible with models generic models.

Resources for the LM

Input data

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    ๐Ÿ–– Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. ๐Ÿ“Š๐Ÿ“ˆ๐ŸŽ‰

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google โค๏ธ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.