Added `--clear-local` to delete local models

01OS/01OS/server/stt/local_service/whisper-rust/.gitignore

@@ -1,10 +0,0 @@
-# Generated by Cargo
-# will have compiled files and executables
-debug/
-target/
-
-# These are backup files generated by rustfmt
-**/*.rs.bk
-
-# MSVC Windows builds of rustc generate these, which store debugging information
-*.pdb

01OS/01OS/server/stt/local_service/whisper-rust/Cargo.toml

@@ -1,14 +0,0 @@
-[package]
-name = "whisper-rust"
-version = "0.1.0"
-edition = "2021"
-
-# See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
-
-[dependencies]
-anyhow = "1.0.79"
-clap = { version = "4.4.18", features = ["derive"] }
-cpal = "0.15.2"
-hound = "3.5.1"
-whisper-rs = "0.10.0"
-whisper-rs-sys = "0.8.0"

01OS/01OS/server/stt/local_service/whisper-rust/README.md

@@ -1,7 +0,0 @@
-# Setup
-
-To rebuild the `whisper-rust` executable, do the following:
-
-1. Install [Rust](https://www.rust-lang.org/tools/install), cmake, and Python dependencies `pip install -r requirements.txt`.
-2. Go to **core/stt** and run `cargo build --release`.
-3. Move the `whisper-rust` executable from target/release to this directory.

01OS/01OS/server/stt/local_service/whisper-rust/src/main.rs

@@ -1,34 +0,0 @@
-mod transcribe;
-
-use clap::Parser;
-use std::path::PathBuf;
-use transcribe::transcribe;
-
-#[derive(Parser, Debug)]
-#[command(author, version, about, long_about = None)]
-struct Args {
-    /// This is the model for Whisper STT
-    #[arg(short, long, value_parser, required = true)]
-    model_path: PathBuf,
-    
-    /// This is the wav audio file that will be converted from speech to text
-    #[arg(short, long, value_parser, required = true)]
-    file_path: Option<PathBuf>,
-}
-
-fn main() {
-
-    let args = Args::parse();
-
-    let file_path = match args.file_path {
-        Some(fp) => fp,
-        None => panic!("No file path provided")
-    };
-
-    let result = transcribe(&args.model_path, &file_path);
-
-    match result {
-        Ok(transcription) => print!("{}", transcription),
-        Err(e) => panic!("Error: {}", e),
-    }
-}

01OS/01OS/server/stt/local_service/whisper-rust/src/transcribe.rs

@@ -1,64 +0,0 @@
-use whisper_rs::{FullParams, SamplingStrategy, WhisperContext, WhisperContextParameters};
-use std::path::PathBuf;
-
-
-/// Transcribes the given audio file using the whisper-rs library.
-///
-/// # Arguments
-/// * `model_path` - Path to Whisper model file
-/// * `file_path` - A string slice that holds the path to the audio file to be transcribed.
-///
-/// # Returns
-///
-/// A Result containing a String with the transcription if successful, or an error message if not.
-pub fn transcribe(model_path: &PathBuf, file_path: &PathBuf) -> Result<String, String> {
-
-    let model_path_str = model_path.to_str().expect("Not valid model path");
-    // Load a context and model
-    let ctx = WhisperContext::new_with_params(
-        model_path_str, // Replace with the actual path to the model
-        WhisperContextParameters::default(),
-    )
-    .map_err(|_| "failed to load model")?;
-
-    // Create a state
-    let mut state = ctx.create_state().map_err(|_| "failed to create state")?;
-
-    // Create a params object
-    // Note that currently the only implemented strategy is Greedy, BeamSearch is a WIP
-    let mut params = FullParams::new(SamplingStrategy::Greedy { best_of: 1 });
-
-    // Edit parameters as needed
-    params.set_n_threads(1); // Set the number of threads to use
-    params.set_translate(true); // Enable translation
-    params.set_language(Some("en")); // Set the language to translate to English
-    // Disable printing to stdout
-    params.set_print_special(false);
-    params.set_print_progress(false);
-    params.set_print_realtime(false);
-    params.set_print_timestamps(false);
-
-    // Load the audio file
-    let audio_data = std::fs::read(file_path)
-        .map_err(|e| format!("failed to read audio file: {}", e))?
-        .chunks_exact(2)
-        .map(|chunk| i16::from_ne_bytes([chunk[0], chunk[1]]))
-        .collect::<Vec<i16>>();
-
-    // Convert the audio data to the required format (16KHz mono i16 samples)
-    let audio_data = whisper_rs::convert_integer_to_float_audio(&audio_data);
-
-    // Run the model
-    state.full(params, &audio_data[..]).map_err(|_| "failed to run model")?;
-
-    // Fetch the results
-    let num_segments = state.full_n_segments().map_err(|_| "failed to get number of segments")?;
-    let mut transcription = String::new();
-    for i in 0..num_segments {
-        let segment = state.full_get_segment_text(i).map_err(|_| "failed to get segment")?;
-        transcription.push_str(&segment);
-        transcription.push('\n');
-    }
-
-    Ok(transcription)
-}

01OS/start.sh

@@ -54,6 +54,16 @@ if [[ "$@" == *"--expose"* ]]; then
     export SERVER_EXPOSE_PUBLICALLY="True"
 fi
 
+# Check if "--clear-local" is passed as an argument
+if [[ "$@" == *"--clear-local"* ]]; then
+    # If "--clear-local" is passed, clear the contents of the folders in script_dir/01OS/server/{tts and stt}/local_service
+    echo "Clearing local services..."
+    rm -rf "$SCRIPT_DIR/01OS/server/tts/local_service"/*
+    rm -rf "$SCRIPT_DIR/01OS/server/stt/local_service"/*
+    echo "Exiting after clearing local services..."
+    exit 0
+fi
+
 ### SETUP
 
 if [[ "$ALL_LOCAL" == "True" ]]; then