File size: 6,979 Bytes
df5d609 4994b71 df5d609 d147321 4994b71 df5d609 4994b71 df5d609 4994b71 df5d609 4994b71 df5d609 4994b71 df5d609 4994b71 df5d609 4994b71 d147321 4994b71 df5d609 4994b71 df5d609 4994b71 df5d609 4994b71 df5d609 4994b71 df5d609 d147321 4994b71 df5d609 4994b71 df5d609 4994b71 0152ed4 4994b71 df5d609 4994b71 df5d609 4994b71 df5d609 4994b71 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 | import json
import sys
import time
from datetime import datetime
from pathlib import Path
import sqlite3
import shutil
from langchain.text_splitter import RecursiveCharacterTextSplitter
from langchain_community.vectorstores import Chroma
from langchain_openai import OpenAIEmbeddings
from tenacity import (
retry,
retry_if_exception,
stop_after_attempt,
wait_exponential
)
from tqdm import tqdm
# Add parent directory to path for imports
sys.path.insert(0, str(Path(__file__).parent.parent))
from config.settings import Settings # noqa: E402
from src.utils.data_loader import LyricsLoader # noqa: E402
class LyricsProcessor:
"""Process lyrics files into embeddings for vector search."""
def __init__(
self,
lyrics_dir: str,
output_dir: str,
batch_size: int = 100
):
self.lyrics_dir = Path(lyrics_dir)
self.output_dir = Path(output_dir)
self.batch_size = batch_size
self.embeddings = OpenAIEmbeddings()
self.collection_name = Settings.CHROMA_COLLECTION_NAME
print(f"Using collection name: {self.collection_name}")
# Configure text splitter for lyrics
self.text_splitter = RecursiveCharacterTextSplitter(
chunk_size=300,
chunk_overlap=75,
separators=["\n\n", "\n", " ", ""],
keep_separator=True
)
# Initialize loader
self.loader = LyricsLoader(lyrics_dir)
# Ensure output directory exists
self.output_dir.mkdir(parents=True, exist_ok=True)
def validate_text_splitter(self):
"""Validate text splitter configuration."""
required_attrs = ['_chunk_size', '_chunk_overlap']
missing_attrs = [
attr for attr in required_attrs
if not hasattr(self.text_splitter, attr)
]
if missing_attrs:
msg = f"Text splitter missing required attributes: {missing_attrs}"
raise AttributeError(msg)
def _is_rate_limit_error(self, exception):
"""Check if exception is a rate limit error."""
return "rate_limit" in str(exception).lower()
@retry(
retry=retry_if_exception(_is_rate_limit_error),
wait=wait_exponential(multiplier=1, min=4, max=10),
stop=stop_after_attempt(3)
)
def _create_embeddings_with_backoff(self, vector_store, batch):
"""Create embeddings with retry logic."""
return vector_store.add_documents(batch)
def process_lyrics(self) -> None:
"""Main processing pipeline for lyrics."""
print("Starting lyrics processing pipeline...")
# Validate configuration before starting
print("Validating configuration...")
self.validate_text_splitter()
# Check for existing collection
chroma_dir = Path(self.output_dir) / "chroma"
if chroma_dir.exists():
sqlite_file = chroma_dir / "chroma.sqlite3"
if sqlite_file.exists():
try:
conn = sqlite3.connect(sqlite_file)
cursor = conn.cursor()
cursor.execute("SELECT name FROM collections WHERE name = ?",
(self.collection_name,))
if cursor.fetchone():
response = input(
f"\nWarning: Collection '{self.collection_name}' already exists.\n"
"Do you want to delete and recreate? (y/N): "
)
if response.lower() != 'y':
print("Aborting.")
return
print("Removing existing collection...")
shutil.rmtree(chroma_dir)
chroma_dir.mkdir(parents=True)
conn.close()
except Exception as e:
print(f"Error checking existing collection: {e}")
print("Continuing with processing...")
# Load all lyrics documents
print("Loading lyrics files...")
documents = self.loader.load_lyrics()
if not documents:
raise ValueError("No valid lyrics files found")
print(f"Loaded {len(documents)} valid lyrics files")
# Split documents
print("Processing documents...")
processed_docs = []
for doc in tqdm(documents, desc="Processing documents"):
splits = self.text_splitter.split_documents([doc])
processed_docs.extend(splits)
n_docs = len(documents)
n_chunks = len(processed_docs)
chunks_msg = f"Created {n_chunks} chunks from {n_docs}"
print(f"{chunks_msg} documents")
# Create embeddings in batches
print("Creating embeddings...")
total_chunks = len(processed_docs)
for i in tqdm(range(0, total_chunks, self.batch_size)):
batch = processed_docs[i:i + self.batch_size]
try:
if i == 0:
# Initialize vector store with first batch
vector_store = Chroma.from_documents(
documents=batch,
embedding=self.embeddings,
persist_directory=str(self.output_dir / "chroma"),
collection_name=self.collection_name
)
else:
# Add subsequent batches
self._create_embeddings_with_backoff(vector_store, batch)
# Rate limit cooldown
time.sleep(2)
except Exception as e:
if self._is_rate_limit_error(e):
print("Rate limit reached. Waiting before retry...")
time.sleep(60)
continue
raise
# Save processing metadata
metadata = {
'processed_at': datetime.now().isoformat(),
'total_documents': len(documents),
'total_chunks': total_chunks,
'chunk_size': getattr(self.text_splitter, '_chunk_size', 300),
'chunk_overlap': getattr(self.text_splitter, '_chunk_overlap', 75)
}
meta_path = self.output_dir / 'processing_metadata.json'
with open(meta_path, 'w') as f:
json.dump(metadata, f, indent=2)
print("Processing complete!")
print(f"Processed {len(documents)} documents into {total_chunks} chunks")
print(f"Embeddings saved to {self.output_dir / 'chroma'}")
if __name__ == "__main__":
processor = LyricsProcessor(
lyrics_dir=str(Settings.LYRICS_DIR),
output_dir=str(Settings.EMBEDDINGS_DIR),
batch_size=100
)
processor.process_lyrics() |