File size: 5,158 Bytes
325b94c | 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 | from __future__ import annotations
from typing import Dict, Any, List
from core.books.storage import (
fetch_raw_docs_for_user,
mark_raw_status,
upsert_document_metadata,
)
from schemas.books.sources_schema import SourceForAgent, DocMetadata
from agents.books.apa_agent import run_metadata_agent
# -------- limits (VERY IMPORTANT) --------
MAX_PAGES = 5
MAX_CHARS_PER_PAGE = 5000
def _build_agent_input(raw: Dict[str, Any]) -> Dict[str, Any]:
pages = raw.get("pages_head") or []
# pages = pages[:MAX_PAGES]
# pages = [p[:MAX_CHARS_PER_PAGE] for p in pages]
return SourceForAgent(
source_url=raw["source_url"],
source_type=raw.get("source_type", "pdf"),
domain=raw.get("domain", ""),
search_title=raw.get("search_title", "") or "",
search_snippet=raw.get("search_snippet", "") or "",
text_pages=pages,
).model_dump()
def process_user_metadata(user_id: str, book_id: str | None = None) -> Dict[str, Any]:
pending = fetch_raw_docs_for_user(
user_id=user_id,
book_id=book_id,
status="pending",
)
total = len(pending)
processed = 0
failed = 0
items: List[Dict[str, Any]] = []
tokens = []
if total == 0:
return {
"user_id": user_id,
"book_id": book_id,
"total": 0,
"processed": 0,
"failed": 0,
"items": [],
"mode": "sequential",
}
for raw in pending:
doc_id = raw["doc_id"]
url = raw["source_url"]
pages_head = raw.get("pages_head") or []
if not pages_head:
mark_raw_status(doc_id, "failed", "no_pages_head")
failed += 1
items.append(
{
"doc_id": doc_id,
"url": url,
"status": "failed",
"reason": "no_pages_head",
}
)
continue
mark_raw_status(doc_id, "processing", "")
try:
agent_in = _build_agent_input(raw)
out, token_usage = run_metadata_agent(agent_in)
except Exception as e:
print(f"⚠️ LLM internal exception for doc_id {doc_id}: {e}")
out = None
# ---- بعد الـ try/except ----
if not isinstance(out, dict):
mark_raw_status(doc_id, "failed", "llm_exception")
failed += 1
items.append(
{
"doc_id": doc_id,
"url": url,
"status": "failed",
"reason": "llm_exception",
}
)
continue
if token_usage:
tokens.append(token_usage)
print(f"Metadata agent output for doc_id {doc_id}: {type(out)} - {out}")
print(f"is_valid: {out.get('is_valid')}")
if not out.get("is_valid"):
reason = out.get("reason", "invalid")
mark_raw_status(doc_id, "failed", reason)
failed += 1
items.append(
{
"doc_id": doc_id,
"url": url,
"status": "failed",
"reason": reason,
}
)
continue
md = out.get("metadata") or {}
if "title" not in md or "authors" not in md:
mark_raw_status(doc_id, "failed", "missing_title_or_authors")
failed += 1
items.append(
{
"doc_id": doc_id,
"url": url,
"status": "failed",
"reason": "missing_title_or_authors",
}
)
continue
try:
doc_md = DocMetadata(
doc_id=doc_id,
title=md["title"],
authors=md["authors"],
year=md.get("year"),
publisher_or_journal=md.get("publisher_or_journal") or "",
normalized_source_type=out.get("normalized_source_type", "pdf"),
apa7=md.get("apa7") or "",
metadata=out,
)
upsert_document_metadata(doc_md)
mark_raw_status(doc_id, "processed", "")
processed += 1
items.append(
{
"doc_id": doc_id,
"url": url,
"status": "processed",
"title": doc_md.title,
}
)
except Exception:
mark_raw_status(doc_id, "failed", "db_failed")
failed += 1
items.append(
{
"doc_id": doc_id,
"url": url,
"status": "failed",
"reason": "db_failed",
}
)
return {
"user_id": user_id,
"book_id": book_id,
"total": total,
"processed": processed,
"failed": failed,
"items": items,
"mode": "sequential",
"tokens": tokens,
}
|