Kapamtalk / evaluate.py
Coco-18's picture
Upload 3 files
0812080 verified
raw
history blame
15.3 kB
# evaluate.py - Handles evaluation and comparing tasks
import os
import glob
import logging
import traceback
import tempfile
import shutil
from difflib import SequenceMatcher
import torch
import torchaudio
from pydub import AudioSegment
from flask import jsonify
from werkzeug.utils import secure_filename
from concurrent.futures import ThreadPoolExecutor
# Import necessary functions from translator.py
from translator import asr_model, asr_processor, LANGUAGE_CODES
# Configure logging
logger = logging.getLogger("speech_api")
def calculate_similarity(text1, text2):
"""Calculate text similarity percentage."""
def clean_text(text):
return text.lower()
clean1 = clean_text(text1)
clean2 = clean_text(text2)
matcher = SequenceMatcher(None, clean1, clean2)
return matcher.ratio() * 100
def init_reference_audio(reference_dir, output_dir):
try:
# Create the output directory first
os.makedirs(output_dir, exist_ok=True)
logger.info(f"πŸ“ Created output directory: {output_dir}")
# Check if the reference audio directory exists in the repository
if os.path.exists(reference_dir):
logger.info(f"βœ… Found reference audio directory: {reference_dir}")
# Log the contents to verify
pattern_dirs = [d for d in os.listdir(reference_dir)
if os.path.isdir(os.path.join(reference_dir, d))]
logger.info(f"πŸ“ Found reference patterns: {pattern_dirs}")
# Check each pattern directory for wav files
for pattern_dir_name in pattern_dirs:
pattern_path = os.path.join(reference_dir, pattern_dir_name)
wav_files = glob.glob(os.path.join(pattern_path, "*.wav"))
logger.info(f"πŸ“ Found {len(wav_files)} wav files in {pattern_dir_name}")
else:
logger.warning(f"⚠️ Reference audio directory not found: {reference_dir}")
# Create the directory if it doesn't exist
os.makedirs(reference_dir, exist_ok=True)
logger.info(f"πŸ“ Created reference audio directory: {reference_dir}")
except Exception as e:
logger.error(f"❌ Failed to set up reference audio directory: {str(e)}")
def handle_upload_reference(request, reference_dir, sample_rate):
"""Handle upload of reference audio files"""
try:
if "audio" not in request.files:
logger.warning("⚠️ Reference upload missing audio file")
return jsonify({"error": "No audio file uploaded"}), 400
reference_word = request.form.get("reference_word", "").strip()
if not reference_word:
logger.warning("⚠️ Reference upload missing reference word")
return jsonify({"error": "No reference word provided"}), 400
# Validate reference word
reference_patterns = [
"mayap_a_abak", "mayap_a_ugtu", "mayap_a_gatpanapun", "mayap_a_bengi",
"komusta_ka", "malaus_ko_pu", "malaus_kayu", "agaganaka_da_ka",
"pagdulapan_da_ka", "kaluguran_da_ka", "dakal_a_salamat", "panapaya_mu_ku"
]
if reference_word not in reference_patterns:
logger.warning(f"⚠️ Invalid reference word: {reference_word}")
return jsonify({"error": f"Invalid reference word. Available: {reference_patterns}"}), 400
# Create directory for reference pattern if it doesn't exist
pattern_dir = os.path.join(reference_dir, reference_word)
os.makedirs(pattern_dir, exist_ok=True)
# Save the reference audio file
audio_file = request.files["audio"]
file_path = os.path.join(pattern_dir, secure_filename(audio_file.filename))
audio_file.save(file_path)
# Convert to WAV if not already in that format
if not file_path.lower().endswith('.wav'):
base_path = os.path.splitext(file_path)[0]
wav_path = f"{base_path}.wav"
try:
audio = AudioSegment.from_file(file_path)
audio = audio.set_frame_rate(sample_rate).set_channels(1)
audio.export(wav_path, format="wav")
# Remove original file if conversion successful
os.unlink(file_path)
file_path = wav_path
except Exception as e:
logger.error(f"❌ Reference audio conversion failed: {str(e)}")
return jsonify({"error": f"Audio conversion failed: {str(e)}"}), 500
logger.info(f"βœ… Reference audio saved successfully for {reference_word}: {file_path}")
# Count how many references we have now
references = glob.glob(os.path.join(pattern_dir, "*.wav"))
return jsonify({
"message": "Reference audio uploaded successfully",
"reference_word": reference_word,
"file": os.path.basename(file_path),
"total_references": len(references)
})
except Exception as e:
logger.error(f"❌ Unhandled exception in reference upload: {str(e)}")
logger.debug(f"Stack trace: {traceback.format_exc()}")
return jsonify({"error": f"Internal server error: {str(e)}"}), 500
def handle_evaluation_request(request, reference_dir, output_dir, sample_rate):
"""Handle pronunciation evaluation requests"""
request_id = f"req-{id(request)}" # Create unique ID for this request
logger.info(f"[{request_id}] πŸ†• Starting new pronunciation evaluation request")
temp_dir = None
if asr_model is None or asr_processor is None:
logger.error(f"[{request_id}] ❌ Evaluation endpoint called but ASR models aren't loaded")
return jsonify({"error": "ASR model not available"}), 503
try:
if "audio" not in request.files:
logger.warning(f"[{request_id}] ⚠️ Evaluation request missing audio file")
return jsonify({"error": "No audio file uploaded"}), 400
audio_file = request.files["audio"]
reference_locator = request.form.get("reference_locator", "").strip()
language = request.form.get("language", "kapampangan").lower()
# Validate reference locator
if not reference_locator:
logger.warning(f"[{request_id}] ⚠️ No reference locator provided")
return jsonify({"error": "Reference locator is required"}), 400
# Construct full reference directory path
reference_dir_path = os.path.join(reference_dir, reference_locator)
logger.info(f"[{request_id}] πŸ“ Reference directory path: {reference_dir_path}")
if not os.path.exists(reference_dir_path):
logger.warning(f"[{request_id}] ⚠️ Reference directory not found: {reference_dir_path}")
return jsonify({"error": f"Reference audio directory not found: {reference_locator}"}), 404
reference_files = glob.glob(os.path.join(reference_dir_path, "*.wav"))
logger.info(f"[{request_id}] πŸ“ Found {len(reference_files)} reference files")
if not reference_files:
logger.warning(f"[{request_id}] ⚠️ No reference audio files found in {reference_dir_path}")
return jsonify({"error": f"No reference audio found for {reference_locator}"}), 404
lang_code = LANGUAGE_CODES.get(language, language)
logger.info(
f"[{request_id}] πŸ”„ Evaluating pronunciation for reference: {reference_locator} with language code: {lang_code}")
# Create a request-specific temp directory to avoid conflicts
temp_dir = os.path.join(output_dir, f"temp_{request_id}")
os.makedirs(temp_dir, exist_ok=True)
# Process user audio
user_audio_path = os.path.join(temp_dir, "user_audio_input.wav")
with open(user_audio_path, 'wb') as f:
f.write(audio_file.read())
try:
logger.info(f"[{request_id}] πŸ”„ Processing user audio file")
audio = AudioSegment.from_file(user_audio_path)
audio = audio.set_frame_rate(sample_rate).set_channels(1)
processed_path = os.path.join(temp_dir, "processed_user_audio.wav")
audio.export(processed_path, format="wav")
user_waveform, sr = torchaudio.load(processed_path)
user_waveform = user_waveform.squeeze().numpy()
logger.info(f"[{request_id}] βœ… User audio processed: {sr}Hz, length: {len(user_waveform)} samples")
user_audio_path = processed_path
except Exception as e:
logger.error(f"[{request_id}] ❌ Audio processing failed: {str(e)}")
return jsonify({"error": f"Audio processing failed: {str(e)}"}), 500
# Transcribe user audio
try:
logger.info(f"[{request_id}] πŸ”„ Transcribing user audio")
inputs = asr_processor(
user_waveform,
sampling_rate=sample_rate,
return_tensors="pt",
language=lang_code
)
inputs = {k: v.to(asr_model.device) for k, v in inputs.items()}
with torch.no_grad():
logits = asr_model(**inputs).logits
ids = torch.argmax(logits, dim=-1)[0]
user_transcription = asr_processor.decode(ids)
logger.info(f"[{request_id}] βœ… User transcription: '{user_transcription}'")
except Exception as e:
logger.error(f"[{request_id}] ❌ ASR inference failed: {str(e)}")
return jsonify({"error": f"ASR inference failed: {str(e)}"}), 500
# Process reference files in batches
batch_size = 2 # Process 2 files at a time - adjust based on your hardware
results = []
best_score = 0
best_reference = None
best_transcription = None
# Use this if you want to limit the number of files to process
max_files_to_check = min(5, len(reference_files)) # Check at most 5 files
reference_files = reference_files[:max_files_to_check]
logger.info(f"[{request_id}] πŸ”„ Processing {len(reference_files)} reference files in batches of {batch_size}")
# Function to process a single reference file
def process_reference_file(ref_file):
ref_filename = os.path.basename(ref_file)
try:
# Load and resample reference audio
ref_waveform, ref_sr = torchaudio.load(ref_file)
if ref_sr != sample_rate:
ref_waveform = torchaudio.transforms.Resample(ref_sr, sample_rate)(ref_waveform)
ref_waveform = ref_waveform.squeeze().numpy()
# Transcribe reference audio
inputs = asr_processor(
ref_waveform,
sampling_rate=sample_rate,
return_tensors="pt",
language=lang_code
)
inputs = {k: v.to(asr_model.device) for k, v in inputs.items()}
with torch.no_grad():
logits = asr_model(**inputs).logits
ids = torch.argmax(logits, dim=-1)[0]
ref_transcription = asr_processor.decode(ids)
# Calculate similarity
similarity = calculate_similarity(user_transcription, ref_transcription)
logger.info(
f"[{request_id}] πŸ“Š Similarity with {ref_filename}: {similarity:.2f}%, transcription: '{ref_transcription}'")
return {
"reference_file": ref_filename,
"reference_text": ref_transcription,
"similarity_score": similarity
}
except Exception as e:
logger.error(f"[{request_id}] ❌ Error processing {ref_filename}: {str(e)}")
return {
"reference_file": ref_filename,
"reference_text": "Error",
"similarity_score": 0,
"error": str(e)
}
# Process files in batches using ThreadPoolExecutor
with ThreadPoolExecutor(max_workers=batch_size) as executor:
batch_results = list(executor.map(process_reference_file, reference_files))
results.extend(batch_results)
# Find the best result
for result in batch_results:
if result["similarity_score"] > best_score:
best_score = result["similarity_score"]
best_reference = result["reference_file"]
best_transcription = result["reference_text"]
# Exit early if we found a very good match (optional)
if best_score > 80.0:
logger.info(f"[{request_id}] 🏁 Found excellent match: {best_score:.2f}%")
break
# Clean up temp files
try:
if temp_dir and os.path.exists(temp_dir):
shutil.rmtree(temp_dir)
logger.debug(f"[{request_id}] 🧹 Cleaned up temporary directory")
except Exception as e:
logger.warning(f"[{request_id}] ⚠️ Failed to clean up temp files: {str(e)}")
# Determine feedback based on score
is_correct = best_score >= 70.0
if best_score >= 90.0:
feedback = "Perfect pronunciation! Excellent job!"
elif best_score >= 80.0:
feedback = "Great pronunciation! Your accent is very good."
elif best_score >= 70.0:
feedback = "Good pronunciation. Keep practicing!"
elif best_score >= 50.0:
feedback = "Fair attempt. Try focusing on the syllables that differ from the sample."
else:
feedback = "Try again. Listen carefully to the sample pronunciation."
logger.info(f"[{request_id}] πŸ“Š Final evaluation results: score={best_score:.2f}%, is_correct={is_correct}")
logger.info(f"[{request_id}] πŸ“ Feedback: '{feedback}'")
logger.info(f"[{request_id}] βœ… Evaluation complete")
# Sort results by score descending
results.sort(key=lambda x: x["similarity_score"], reverse=True)
return jsonify({
"is_correct": is_correct,
"score": best_score,
"feedback": feedback,
"user_transcription": user_transcription,
"best_reference_transcription": best_transcription,
"reference_locator": reference_locator,
"details": results
})
except Exception as e:
logger.error(f"[{request_id}] ❌ Unhandled exception in evaluation endpoint: {str(e)}")
logger.debug(f"[{request_id}] Stack trace: {traceback.format_exc()}")
# Clean up on error
try:
if temp_dir and os.path.exists(temp_dir):
shutil.rmtree(temp_dir)
except:
pass
return jsonify({"error": f"Internal server error: {str(e)}"}), 500