Spaces:
Running
Running
File size: 10,493 Bytes
54437b9 26d4279 54437b9 26d4279 54437b9 26d4279 54437b9 26d4279 54437b9 26d4279 54437b9 26d4279 54437b9 | 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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 | """Inkference FastAPI app.
Endpoints:
GET /api/health
GET /api/documents
POST /api/documents
GET /api/documents/{id}
POST /api/documents/{id}/pages upload scans -> background ingest job
GET /api/documents/{id}/pages/{n} transcription (lines + words + confidence)
GET /api/documents/{id}/pages/{n}/image
POST /api/documents/{id}/ask RAG answer + source pages (fast path)
POST /api/documents/{id}/agent LangGraph research agent (multi-step + memory)
DEL /api/documents/{id}/agent/threads/{thread_id} forget one conversation
GET /api/jobs/{id} ingestion progress
"""
from __future__ import annotations
import logging
import os
import re
import unicodedata
from pathlib import Path
from fastapi import FastAPI, File, HTTPException, UploadFile
from fastapi.middleware.cors import CORSMiddleware
from fastapi.responses import FileResponse, RedirectResponse
from fastapi.staticfiles import StaticFiles
from pydantic import BaseModel
from ..config import (
FRONTEND_DIR,
IMAGES_BASE_URL,
IMAGES_ROOT,
agent as agent_cfg,
correction as correction_cfg,
htr as htr_cfg,
rag as rag_cfg,
)
from ..rag.answer import answer_question
from . import services
def _setup_logging() -> None:
"""Route all `inkference.*` loggers to the console at INKFERENCE_LOG_LEVEL
(default INFO; set DEBUG for verbose). Independent of uvicorn's own loggers."""
level = os.getenv("INKFERENCE_LOG_LEVEL", "INFO").upper()
lg = logging.getLogger("inkference")
lg.setLevel(level)
if not lg.handlers:
h = logging.StreamHandler()
h.setFormatter(logging.Formatter(
"%(asctime)s [%(levelname)s] %(name)s: %(message)s", "%H:%M:%S"))
lg.addHandler(h)
lg.propagate = False
_setup_logging()
logger = logging.getLogger("inkference.api")
app = FastAPI(title="Inkference", version="0.1.0")
app.add_middleware(
CORSMiddleware,
allow_origins=os.getenv("CORS_ORIGINS", "*").split(","),
allow_methods=["*"],
allow_headers=["*"],
)
@app.middleware("http")
async def _no_cache_frontend(request, call_next):
"""Tell browsers to revalidate the static frontend so edits (js/css/html)
always load fresh instead of serving a stale cached bundle."""
response = await call_next(request)
if not request.url.path.startswith("/api/"):
response.headers["Cache-Control"] = "no-cache, no-store, must-revalidate"
return response
# --------------------------------------------------------------------------- #
# request models
# --------------------------------------------------------------------------- #
class CreateDocument(BaseModel):
title: str
slug: str | None = None
subtitle: str | None = None
class AskRequest(BaseModel):
question: str
top_k: int | None = None
persona: str | None = None # "cook" -> answer in character as Captain Cook
class AgentAskRequest(BaseModel):
question: str
# Conversation id, minted by the client. Omit for a one-off (unremembered) turn.
thread_id: str | None = None
persona: str | None = None
max_steps: int | None = None
def _slugify(text: str) -> str:
text = unicodedata.normalize("NFKD", text).encode("ascii", "ignore").decode()
return re.sub(r"[^a-z0-9]+", "-", text.lower()).strip("-") or "document"
# --------------------------------------------------------------------------- #
# documents
# --------------------------------------------------------------------------- #
@app.get("/api/health")
def health() -> dict:
return {"status": "ok", "trocr_model": htr_cfg.trocr_model_id,
"llm_provider": rag_cfg.llm_provider,
"llm_configured": bool(rag_cfg.llm_api_key),
"correction_enabled": correction_cfg.enabled,
"correction_backend": correction_cfg.backend,
"correction_model": (correction_cfg.api_model if correction_cfg.backend == "api"
else correction_cfg.model_id),
"agent_enabled": agent_cfg.enabled,
"agent_max_steps": agent_cfg.max_steps}
@app.get("/api/documents")
def list_documents() -> list[dict]:
return services.get_store().list_documents()
@app.post("/api/documents")
def create_document(body: CreateDocument) -> dict:
store = services.get_store()
slug = body.slug or _slugify(body.title)
if store.get_document_by_slug(slug):
raise HTTPException(409, f"document with slug '{slug}' already exists")
doc_id = store.create_document(title=body.title, slug=slug, subtitle=body.subtitle)
return {"id": doc_id, "slug": slug}
@app.get("/api/documents/{doc_id}")
def get_document(doc_id: int) -> dict:
doc = services.get_store().get_document(doc_id)
if not doc:
raise HTTPException(404, "document not found")
return doc
# --------------------------------------------------------------------------- #
# pages: upload + ingest
# --------------------------------------------------------------------------- #
@app.post("/api/documents/{doc_id}/pages")
async def upload_pages(doc_id: int, files: list[UploadFile] = File(...)) -> dict:
store = services.get_store()
doc = store.get_document(doc_id)
if not doc:
raise HTTPException(404, "document not found")
assets = store.cfg.assets_dir / doc["slug"]
assets.mkdir(parents=True, exist_ok=True)
start = doc["page_count"] + 1
specs: list[tuple[int, int, str]] = []
for offset, upload in enumerate(files):
page_number = start + offset
ext = Path(upload.filename or "").suffix.lower() or ".png"
dest = assets / f"page_{page_number:04d}{ext}"
dest.write_bytes(await upload.read())
page_id = store.add_page(doc_id, page_number, image_path=str(dest))
specs.append((page_id, page_number, str(dest)))
job_id = store.create_job(doc_id, total_pages=len(specs))
services.submit_ingest(doc_id, specs, job_id)
return {"job_id": job_id, "pages": [s[1] for s in specs]}
@app.get("/api/jobs/{job_id}")
def get_job(job_id: int) -> dict:
job = services.get_store().get_job(job_id)
if not job:
raise HTTPException(404, "job not found")
return job
# --------------------------------------------------------------------------- #
# pages: read
# --------------------------------------------------------------------------- #
@app.get("/api/documents/{doc_id}/pages/{page_number}")
def get_page(doc_id: int, page_number: int) -> dict:
page = services.get_store().get_page(doc_id, page_number)
if not page:
raise HTTPException(404, "page not found")
page.pop("image_path", None) # internal path; image served via dedicated route
return page
@app.get("/api/documents/{doc_id}/pages/{page_number}/image")
def get_page_image(doc_id: int, page_number: int):
path = services.get_store().get_page_image_path(doc_id, page_number)
if not path:
raise HTTPException(404, "page image not found")
p = Path(path)
# Page NUMBERS are reused across uploads/reseeds, so /pages/{n}/image can map to
# different bytes over time. Forbid browser caching so a reused number never serves
# a stale scan (the "correct transcript but wrong old page image" bug).
no_cache = {"Cache-Control": "no-cache, no-store, must-revalidate"}
# Remote images (deployment): redirect relative keys to a CDN/dataset base URL
# so large image sets don't need to be baked into the app.
if IMAGES_BASE_URL and not p.is_absolute():
return RedirectResponse(f"{IMAGES_BASE_URL.rstrip('/')}/{path}", headers=no_cache)
# Local: absolute path (self-contained seeds) or relative key under IMAGES_ROOT.
for candidate in (p, (IMAGES_ROOT / path) if IMAGES_ROOT else None):
if candidate and candidate.exists():
return FileResponse(candidate, headers=no_cache)
raise HTTPException(404, "page image not found")
# --------------------------------------------------------------------------- #
# ask the archive (RAG)
# --------------------------------------------------------------------------- #
@app.post("/api/documents/{doc_id}/ask")
def ask(doc_id: int, body: AskRequest) -> dict:
store = services.get_store()
if not store.get_document(doc_id):
raise HTTPException(404, "document not found")
index = services.get_index()
if not index.exists(doc_id):
logger.info("building RAG index for doc %s", doc_id)
index.build_from_store(doc_id, store)
logger.info("ask doc=%s persona=%s q=%r", doc_id, body.persona, body.question[:100])
ans = answer_question(doc_id, body.question, index, top_k=body.top_k, persona=body.persona)
logger.info("ask doc=%s -> sources=%s", doc_id, ans.source_pages)
return ans.to_dict()
# --------------------------------------------------------------------------- #
# ask the archive (LangGraph agent)
# --------------------------------------------------------------------------- #
# /ask above stays the fast path (~1 LLM call). The agent adds a tool-using loop
# for the questions /ask cannot serve: narrative "what happened next" questions
# that need pages read in order, and follow-ups that need conversation memory.
@app.post("/api/documents/{doc_id}/agent")
def ask_agent(doc_id: int, body: AgentAskRequest) -> dict:
if not agent_cfg.enabled:
raise HTTPException(503, "agent is disabled (AGENT_ENABLED=false)")
if not services.get_store().get_document(doc_id):
raise HTTPException(404, "document not found")
ans = services.run_agent(
doc_id, body.question,
thread_id=body.thread_id, persona=body.persona, max_steps=body.max_steps,
)
return ans.to_dict()
@app.delete("/api/documents/{doc_id}/agent/threads/{thread_id}")
def delete_agent_thread(doc_id: int, thread_id: str) -> dict:
"""Forget one conversation ("New conversation" in the UI).
Doc-scoped because runner.run_agent namespaces thread ids as "{doc_id}:{id}".
"""
from ..agent.checkpoint import delete_thread
return {"deleted": delete_thread(f"{doc_id}:{thread_id}"), "thread_id": thread_id}
# --------------------------------------------------------------------------- #
# static frontend (mounted last so /api/* wins)
# --------------------------------------------------------------------------- #
if FRONTEND_DIR.is_dir():
app.mount("/", StaticFiles(directory=str(FRONTEND_DIR), html=True), name="frontend")
|