Spaces:
Running
Running
AuthorBot commited on
Commit Β·
745bdc2
1
Parent(s): a806a0b
Fix ingest.py: use correct service signatures (chunk_document, summarize_book, file_path args)
Browse files- app/api/ingest.py +50 -35
app/api/ingest.py
CHANGED
|
@@ -4,23 +4,21 @@ Routes (per implementation plan):
|
|
| 4 |
POST /api/admin/{author_slug}/books/upload β Upload a book file
|
| 5 |
GET /api/admin/{author_slug}/books/upload/progress β SSE ingestion progress
|
| 6 |
|
| 7 |
-
|
| 8 |
-
|
| 9 |
"""
|
| 10 |
|
| 11 |
import asyncio
|
| 12 |
import json
|
|
|
|
|
|
|
|
|
|
| 13 |
|
| 14 |
from fastapi import APIRouter, Depends, File, UploadFile, HTTPException
|
| 15 |
from fastapi.responses import StreamingResponse
|
| 16 |
from sqlalchemy.ext.asyncio import AsyncSession
|
| 17 |
|
| 18 |
from app.dependencies import get_db, get_current_user, get_redis
|
| 19 |
-
from app.services.file_validator import validate_file
|
| 20 |
-
from app.services.parser import parse_document
|
| 21 |
-
from app.services.chunker import chunk_text
|
| 22 |
-
from app.services.embeddings import embed_and_store
|
| 23 |
-
from app.services.summarizer import generate_summary
|
| 24 |
from app.repositories.book_repo import BookRepository
|
| 25 |
from app.repositories.document_repo import DocumentRepository
|
| 26 |
from app.config import get_settings
|
|
@@ -41,42 +39,42 @@ async def upload_book(
|
|
| 41 |
db: AsyncSession = Depends(get_db),
|
| 42 |
redis=Depends(get_redis),
|
| 43 |
):
|
| 44 |
-
"""Upload a book file (PDF/EPUB/DOCX/TXT) and start async ingestion.
|
| 45 |
-
|
| 46 |
-
Responds immediately with book_id. Use /progress SSE to track ingestion.
|
| 47 |
-
"""
|
| 48 |
-
# Validate file (magic bytes, not just extension)
|
| 49 |
contents = await file.read()
|
| 50 |
-
|
| 51 |
|
| 52 |
book_repo = BookRepository(db)
|
| 53 |
doc_repo = DocumentRepository(db)
|
| 54 |
|
| 55 |
book = await book_repo.create(
|
| 56 |
author_id=current_user.id,
|
| 57 |
-
title=title or
|
| 58 |
genre=genre,
|
| 59 |
price=price,
|
| 60 |
buy_url=buy_url,
|
| 61 |
status="processing",
|
| 62 |
)
|
| 63 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 64 |
doc = await doc_repo.create(
|
| 65 |
book_id=book.id,
|
| 66 |
author_id=current_user.id,
|
| 67 |
-
filename=
|
| 68 |
-
file_extension=
|
| 69 |
file_size_bytes=len(contents),
|
| 70 |
)
|
| 71 |
await db.commit()
|
| 72 |
|
| 73 |
-
# Fire async ingestion in background
|
| 74 |
asyncio.create_task(
|
| 75 |
_run_ingestion(
|
| 76 |
-
|
| 77 |
-
ext=ext,
|
| 78 |
book_id=book.id,
|
| 79 |
-
doc_id=doc.id,
|
| 80 |
author_id=current_user.id,
|
| 81 |
title=book.title,
|
| 82 |
redis=redis,
|
|
@@ -94,7 +92,7 @@ async def ingestion_progress(
|
|
| 94 |
current_user=Depends(get_current_user),
|
| 95 |
redis=Depends(get_redis),
|
| 96 |
):
|
| 97 |
-
"""SSE stream for real-time ingestion progress
|
| 98 |
async def event_stream():
|
| 99 |
channel = f"ingestion:{current_user.id}:{book_id}"
|
| 100 |
pubsub = redis.pubsub()
|
|
@@ -117,36 +115,46 @@ async def ingestion_progress(
|
|
| 117 |
|
| 118 |
|
| 119 |
async def _publish(redis, channel: str, stage: str, progress: int, detail: str = "") -> None:
|
| 120 |
-
|
| 121 |
-
|
|
|
|
|
|
|
|
|
|
| 122 |
|
| 123 |
|
| 124 |
async def _run_ingestion(
|
| 125 |
-
|
| 126 |
-
ext: str,
|
| 127 |
book_id: str,
|
| 128 |
-
doc_id: str,
|
| 129 |
author_id: str,
|
| 130 |
title: str,
|
| 131 |
redis,
|
| 132 |
db: AsyncSession,
|
| 133 |
) -> None:
|
| 134 |
-
"""Full async ingestion: parse β chunk β embed β summarize."""
|
| 135 |
channel = f"ingestion:{author_id}:{book_id}"
|
| 136 |
book_repo = BookRepository(db)
|
| 137 |
|
| 138 |
try:
|
| 139 |
-
await _publish(redis, channel, "
|
| 140 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 141 |
|
| 142 |
-
await _publish(redis, channel, "chunking",
|
| 143 |
-
|
|
|
|
| 144 |
|
| 145 |
-
await _publish(redis, channel, "embedding",
|
|
|
|
| 146 |
collection_name = await embed_and_store(chunks, author_id, book_id, title)
|
| 147 |
|
| 148 |
-
await _publish(redis, channel, "summarizing",
|
| 149 |
-
|
|
|
|
| 150 |
|
| 151 |
await book_repo.update(book_id, {
|
| 152 |
"status": "active",
|
|
@@ -155,10 +163,17 @@ async def _run_ingestion(
|
|
| 155 |
"chunk_count": len(chunks),
|
| 156 |
})
|
| 157 |
await db.commit()
|
| 158 |
-
|
| 159 |
await _publish(redis, channel, "complete", 100)
|
| 160 |
|
| 161 |
except Exception as e:
|
| 162 |
await book_repo.update(book_id, {"status": "failed"})
|
| 163 |
await db.commit()
|
| 164 |
await _publish(redis, channel, "error", 0, str(e))
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 4 |
POST /api/admin/{author_slug}/books/upload β Upload a book file
|
| 5 |
GET /api/admin/{author_slug}/books/upload/progress β SSE ingestion progress
|
| 6 |
|
| 7 |
+
Saves uploaded file to disk first, then passes the path to services
|
| 8 |
+
that expect file_path (not raw bytes).
|
| 9 |
"""
|
| 10 |
|
| 11 |
import asyncio
|
| 12 |
import json
|
| 13 |
+
import os
|
| 14 |
+
import tempfile
|
| 15 |
+
from pathlib import Path
|
| 16 |
|
| 17 |
from fastapi import APIRouter, Depends, File, UploadFile, HTTPException
|
| 18 |
from fastapi.responses import StreamingResponse
|
| 19 |
from sqlalchemy.ext.asyncio import AsyncSession
|
| 20 |
|
| 21 |
from app.dependencies import get_db, get_current_user, get_redis
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 22 |
from app.repositories.book_repo import BookRepository
|
| 23 |
from app.repositories.document_repo import DocumentRepository
|
| 24 |
from app.config import get_settings
|
|
|
|
| 39 |
db: AsyncSession = Depends(get_db),
|
| 40 |
redis=Depends(get_redis),
|
| 41 |
):
|
| 42 |
+
"""Upload a book file (PDF/EPUB/DOCX/TXT) and start async ingestion."""
|
|
|
|
|
|
|
|
|
|
|
|
|
| 43 |
contents = await file.read()
|
| 44 |
+
filename = file.filename or "upload"
|
| 45 |
|
| 46 |
book_repo = BookRepository(db)
|
| 47 |
doc_repo = DocumentRepository(db)
|
| 48 |
|
| 49 |
book = await book_repo.create(
|
| 50 |
author_id=current_user.id,
|
| 51 |
+
title=title or filename.rsplit(".", 1)[0],
|
| 52 |
genre=genre,
|
| 53 |
price=price,
|
| 54 |
buy_url=buy_url,
|
| 55 |
status="processing",
|
| 56 |
)
|
| 57 |
|
| 58 |
+
# Save to temp dir so file-path-based services can access it
|
| 59 |
+
upload_dir = Path(cfg.UPLOAD_STORAGE_PATH)
|
| 60 |
+
upload_dir.mkdir(parents=True, exist_ok=True)
|
| 61 |
+
suffix = Path(filename).suffix or ".bin"
|
| 62 |
+
tmp_path = upload_dir / f"{book.id}{suffix}"
|
| 63 |
+
tmp_path.write_bytes(contents)
|
| 64 |
+
|
| 65 |
doc = await doc_repo.create(
|
| 66 |
book_id=book.id,
|
| 67 |
author_id=current_user.id,
|
| 68 |
+
filename=filename,
|
| 69 |
+
file_extension=suffix.lstrip("."),
|
| 70 |
file_size_bytes=len(contents),
|
| 71 |
)
|
| 72 |
await db.commit()
|
| 73 |
|
|
|
|
| 74 |
asyncio.create_task(
|
| 75 |
_run_ingestion(
|
| 76 |
+
file_path=str(tmp_path),
|
|
|
|
| 77 |
book_id=book.id,
|
|
|
|
| 78 |
author_id=current_user.id,
|
| 79 |
title=book.title,
|
| 80 |
redis=redis,
|
|
|
|
| 92 |
current_user=Depends(get_current_user),
|
| 93 |
redis=Depends(get_redis),
|
| 94 |
):
|
| 95 |
+
"""SSE stream for real-time ingestion progress."""
|
| 96 |
async def event_stream():
|
| 97 |
channel = f"ingestion:{current_user.id}:{book_id}"
|
| 98 |
pubsub = redis.pubsub()
|
|
|
|
| 115 |
|
| 116 |
|
| 117 |
async def _publish(redis, channel: str, stage: str, progress: int, detail: str = "") -> None:
|
| 118 |
+
if redis:
|
| 119 |
+
try:
|
| 120 |
+
await redis.publish(channel, json.dumps({"stage": stage, "progress": progress, "detail": detail}))
|
| 121 |
+
except Exception:
|
| 122 |
+
pass
|
| 123 |
|
| 124 |
|
| 125 |
async def _run_ingestion(
|
| 126 |
+
file_path: str,
|
|
|
|
| 127 |
book_id: str,
|
|
|
|
| 128 |
author_id: str,
|
| 129 |
title: str,
|
| 130 |
redis,
|
| 131 |
db: AsyncSession,
|
| 132 |
) -> None:
|
| 133 |
+
"""Full async ingestion: validate β parse β chunk β embed β summarize."""
|
| 134 |
channel = f"ingestion:{author_id}:{book_id}"
|
| 135 |
book_repo = BookRepository(db)
|
| 136 |
|
| 137 |
try:
|
| 138 |
+
await _publish(redis, channel, "validating", 5)
|
| 139 |
+
from app.services.file_validator import validate_file
|
| 140 |
+
ext = validate_file(file_path)
|
| 141 |
+
|
| 142 |
+
await _publish(redis, channel, "parsing", 20)
|
| 143 |
+
from app.services.parser import parse_document
|
| 144 |
+
parse_result = parse_document(file_path, ext)
|
| 145 |
+
text = parse_result.text if hasattr(parse_result, "text") else str(parse_result)
|
| 146 |
|
| 147 |
+
await _publish(redis, channel, "chunking", 40)
|
| 148 |
+
from app.services.chunker import chunk_document
|
| 149 |
+
chunks = chunk_document(text)
|
| 150 |
|
| 151 |
+
await _publish(redis, channel, "embedding", 60)
|
| 152 |
+
from app.services.embeddings import embed_and_store
|
| 153 |
collection_name = await embed_and_store(chunks, author_id, book_id, title)
|
| 154 |
|
| 155 |
+
await _publish(redis, channel, "summarizing", 85)
|
| 156 |
+
from app.services.summarizer import summarize_book
|
| 157 |
+
summary = await summarize_book(text[:3000])
|
| 158 |
|
| 159 |
await book_repo.update(book_id, {
|
| 160 |
"status": "active",
|
|
|
|
| 163 |
"chunk_count": len(chunks),
|
| 164 |
})
|
| 165 |
await db.commit()
|
|
|
|
| 166 |
await _publish(redis, channel, "complete", 100)
|
| 167 |
|
| 168 |
except Exception as e:
|
| 169 |
await book_repo.update(book_id, {"status": "failed"})
|
| 170 |
await db.commit()
|
| 171 |
await _publish(redis, channel, "error", 0, str(e))
|
| 172 |
+
|
| 173 |
+
finally:
|
| 174 |
+
# Clean up temp file
|
| 175 |
+
try:
|
| 176 |
+
if os.path.exists(file_path):
|
| 177 |
+
os.remove(file_path)
|
| 178 |
+
except Exception:
|
| 179 |
+
pass
|