Spaces:
Sleeping
Sleeping
File size: 7,876 Bytes
e53b66a 99da31c e53b66a 99da31c e53b66a 99da31c e53b66a 99da31c e53b66a 99da31c e53b66a 99da31c e53b66a 99da31c e53b66a 99da31c e53b66a 99da31c e53b66a | 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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 | #!/usr/bin/env python3
"""Ingest raw JSONL documents into the RAG stores.
Expected input format (one JSON object per line):
- Required: text
- Optional: doc_id, title
This script is intentionally separate from the pre-chunked pipeline:
- pipeline_cli.py expects already chunked rows with chunk_id/chunk_index metadata
- this script accepts raw documents, chunks them locally, and ingests each chunk
using the shared ChunkIngestService write contract.
"""
from __future__ import annotations
import argparse
import hashlib
import json
import os
import sys
from pathlib import Path
from typing import Any, Iterator
# Allow running from project root without installing the package.
sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
from langchain_text_splitters import RecursiveCharacterTextSplitter
from src.core.config import settings
from src.core.dependencies import get_document_repository, get_vector_db_manager
from src.services.chunk_ingest_service import ChunkIngestService
def parse_args() -> argparse.Namespace:
parser = argparse.ArgumentParser(description="Ingest raw JSONL documents using ChunkIngestService.")
parser.add_argument("--input", required=True, help="Path to raw documents JSONL")
parser.add_argument("--limit", type=int, default=0, help="Optional max document count (0=all)")
parser.add_argument(
"--reset-chroma-collection",
action="store_true",
help="Delete and recreate target Chroma collection before ingest.",
)
parser.add_argument(
"--reset-document-store",
action="store_true",
help="Delete all rows from SQLite document store before ingest.",
)
parser.add_argument(
"--reset-all",
action="store_true",
help="Reset both Chroma and SQLite stores before ingest.",
)
parser.add_argument(
"--fail-on-skip",
action="store_true",
help="Return non-zero when any row is skipped (invalid JSON, bad type, empty text).",
)
return parser.parse_args()
def build_chunk_id(source_name: str, chunk_index: int, chunk_text: str) -> str:
digest = hashlib.sha256(chunk_text.encode("utf-8")).hexdigest()[:16]
return f"{source_name}_chunk_{chunk_index}_{digest}"
def build_chunk_metadata(row: dict[str, Any], chunk_index: int, chunk_text: str) -> dict:
return {
"doc_id": row.get("doc_id", row.get("source_name", "unknown")),
"chunk_index": chunk_index,
"title": row.get("title"),
"url": row.get("url"),
"source": row.get("source"),
"authority": row.get("authority"),
"jurisdiction": row.get("jurisdiction"),
"language": row.get("language"),
"topic": row.get("topic"),
"source_family": row.get("source_family"),
"source_type": row.get("source_type"),
"legal_weight": row.get("legal_weight"),
"chunk_word_count": len(chunk_text.split()),
"chunk_char_count": len(chunk_text),
"content_hash": row.get("content_hash"),
}
def _iter_rows(path: Path) -> Iterator[tuple[int, dict[str, Any] | None, str | None]]:
with path.open("r", encoding="utf-8", errors="replace") as f:
for lineno, line in enumerate(f, start=1):
line = line.strip()
if not line:
continue
try:
row = json.loads(line)
except json.JSONDecodeError as exc:
yield lineno, None, f"invalid JSON ({exc})"
continue
if not isinstance(row, dict):
yield lineno, None, f"JSON root must be object/dict, got {type(row).__name__}"
continue
yield lineno, row, None
def _source_name_for_row(row: dict[str, Any], lineno: int) -> str:
source_name = (
str(row.get("doc_id") or "").strip()
or str(row.get("title") or "").strip()
or f"line_{lineno}"
)
return " ".join(source_name.split())[:200] or f"line_{lineno}"
def _reset_chroma(vector_db) -> None:
collection_name = vector_db.collection_name
print(f"resetting_chroma_collection: {collection_name}")
try:
vector_db.client.delete_collection(name=collection_name)
except Exception:
pass
vector_db.collection = vector_db._init_collection()
def _reset_document_store(document_repo) -> None:
document_repo.delete_all_chunks()
print("reset_document_store_deleted_rows: all")
def main() -> int:
args = parse_args()
input_path = Path(args.input)
if not input_path.exists():
print(f"ERROR: input file not found: {input_path}")
return 1
if not input_path.is_file():
print(f"ERROR: input path is not a file: {input_path}")
return 1
try:
vector_db = get_vector_db_manager()
document_repo = get_document_repository()
ingest_service = ChunkIngestService(
document_repo=document_repo,
vector_db=vector_db,
)
except Exception as exc:
print(f"ERROR: dependency initialization failed: {exc}")
return 1
splitter = RecursiveCharacterTextSplitter(
chunk_size=settings.CHUNK_SIZE,
chunk_overlap=settings.CHUNK_OVERLAP,
length_function=len,
is_separator_regex=False,
)
if args.reset_all or args.reset_chroma_collection:
_reset_chroma(vector_db)
if args.reset_all or args.reset_document_store:
_reset_document_store(document_repo)
processed_docs = 0
skipped_docs = 0
failed_docs = 0
total_chunks = 0
seen_docs = 0
try:
for lineno, row, parse_error in _iter_rows(input_path):
if parse_error:
skipped_docs += 1
print(f"SKIP line={lineno}: {parse_error}")
continue
assert row is not None
text = (row.get("text") or "").strip()
if not text:
skipped_docs += 1
print(f"SKIP line={lineno}: missing/empty 'text'")
continue
seen_docs += 1
if args.limit and seen_docs > args.limit:
break
source_name = _source_name_for_row(row, lineno)
try:
chunks = splitter.split_text(text)
chunk_count = 0
for chunk_index, chunk_text in enumerate(chunks):
chunk_id = build_chunk_id(source_name, chunk_index, chunk_text)
metadata = build_chunk_metadata(row, chunk_index, chunk_text)
result = ingest_service.ingest_chunk(
chunk_id=chunk_id,
text=chunk_text,
metadata=metadata,
)
if not (result.document_store_written and result.chroma_written):
raise RuntimeError(f"Ingest failed for chunk_id={chunk_id}")
chunk_count += 1
processed_docs += 1
total_chunks += chunk_count
print(f"OK line={lineno} source='{source_name}' chunks={chunk_count}")
except Exception as exc:
failed_docs += 1
print(f"FAIL line={lineno} source='{source_name}': {exc}")
except KeyboardInterrupt:
print("\nInterrupted by user.")
return 130
print("\nINGEST SUMMARY")
print(f"processed_docs: {processed_docs}")
print(f"failed_docs: {failed_docs}")
print(f"skipped_docs: {skipped_docs}")
print(f"total_chunks: {total_chunks}")
print(f"chroma_ids: {len(vector_db.list_ids())}")
print(f"doc_ids: {len(document_repo.list_chunk_ids())}")
if failed_docs > 0:
return 2
if args.fail_on_skip and skipped_docs > 0:
return 3
return 0
if __name__ == "__main__":
raise SystemExit(main())
|