Spaces:
Sleeping
Validate patent_id shape before hitting Google Patents/OPS
Browse filespatent_id (both the /scrap and /ops path params, and scrap_patents_bulk's
list) was an unvalidated string interpolated straight into a Google
Patents/OPS URL. Not a host-controllable SSRF (the host is a fixed
literal), but garbage input still reached the upstream API and came back
as a confusing 404/502 instead of a clear client error.
- serp.py: split PATENT_ID_REGEX into an unanchored PATENT_ID_CORE (used
to find an id inside scraped text) and the existing bounded regex built
from it, so there's one definition of "what a patent id looks like"
instead of two that could drift.
- app.py: added a PatentId = Annotated[str, StringConstraints(pattern=...)]
built from that same core pattern, applied to both path params and
ScrapPatentsRequest.patent_ids. Malformed input now gets a 422 with a
clear message instead of reaching the upstream API.
Wrote the new tests first (expecting 422 on malformed input), confirmed
they failed against the unvalidated endpoints, then added the validation.
85 tests pass.
- app.py +13 -7
- serp.py +8 -2
- tests/test_app.py +37 -0
|
@@ -3,22 +3,28 @@ import logging
|
|
| 3 |
import os
|
| 4 |
import secrets
|
| 5 |
from contextlib import asynccontextmanager
|
| 6 |
-
from typing import Optional
|
| 7 |
from fastapi import APIRouter, FastAPI, HTTPException, Request
|
| 8 |
from fastapi.responses import JSONResponse
|
| 9 |
from fastapi.routing import APIRouter
|
| 10 |
import httpx
|
| 11 |
from httpx import HTTPStatusError
|
| 12 |
-
from pydantic import BaseModel, Field
|
| 13 |
from playwright.async_api import async_playwright, Browser, BrowserContext, Page
|
| 14 |
import uvicorn
|
| 15 |
|
| 16 |
from scrap import PatentScrapBulkResponse, PatentScrapResult, scrap_patent_async, scrap_patent_bulk_async
|
| 17 |
-
from serp import SerpQuery, SerpResults, query_arxiv, query_bing_search, query_brave_search, query_ddg_search, query_google_patents, query_google_scholar
|
| 18 |
from ops import OPSBulkResponse, OPSNotConfigured, ops_scrap_patent, ops_scrap_patent_bulk, ops_search, token_manager as ops_token_manager
|
| 19 |
from utils import log_gathered_exceptions
|
| 20 |
from mcp_server import mount_mcp_server
|
| 21 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 22 |
logging.basicConfig(
|
| 23 |
level=logging.INFO,
|
| 24 |
format='[%(asctime)s][%(levelname)s][%(filename)s:%(lineno)d]: %(message)s',
|
|
@@ -235,7 +241,7 @@ async def search(params: SerpQuery):
|
|
| 235 |
|
| 236 |
|
| 237 |
@scrap_router.get("/scrap_patent/{patent_id}")
|
| 238 |
-
async def scrap_patent(patent_id:
|
| 239 |
"""Scraps the specified patent from Google Patents.
|
| 240 |
|
| 241 |
Falls back to the EPO OPS API (which covers patents missing from Google
|
|
@@ -274,8 +280,8 @@ async def scrap_patent(patent_id: str) -> PatentScrapResult:
|
|
| 274 |
|
| 275 |
class ScrapPatentsRequest(BaseModel):
|
| 276 |
"""Request model for scrapping multiple patents."""
|
| 277 |
-
patent_ids: list[
|
| 278 |
-
|
| 279 |
|
| 280 |
|
| 281 |
@scrap_router.post("/scrap_patents_bulk", response_model=PatentScrapBulkResponse)
|
|
@@ -296,7 +302,7 @@ async def ops_keyword_search(params: SerpQuery) -> SerpResults:
|
|
| 296 |
|
| 297 |
|
| 298 |
@ops_router.get("/scrap_patent/{patent_id}")
|
| 299 |
-
async def ops_get_patent(patent_id:
|
| 300 |
"""Retrieves a patent (biblio + abstract + claims + description) via EPO OPS."""
|
| 301 |
if not ops_token_manager.configured:
|
| 302 |
raise HTTPException(
|
|
|
|
| 3 |
import os
|
| 4 |
import secrets
|
| 5 |
from contextlib import asynccontextmanager
|
| 6 |
+
from typing import Annotated, Optional
|
| 7 |
from fastapi import APIRouter, FastAPI, HTTPException, Request
|
| 8 |
from fastapi.responses import JSONResponse
|
| 9 |
from fastapi.routing import APIRouter
|
| 10 |
import httpx
|
| 11 |
from httpx import HTTPStatusError
|
| 12 |
+
from pydantic import BaseModel, Field, StringConstraints
|
| 13 |
from playwright.async_api import async_playwright, Browser, BrowserContext, Page
|
| 14 |
import uvicorn
|
| 15 |
|
| 16 |
from scrap import PatentScrapBulkResponse, PatentScrapResult, scrap_patent_async, scrap_patent_bulk_async
|
| 17 |
+
from serp import PATENT_ID_CORE, SerpQuery, SerpResults, query_arxiv, query_bing_search, query_brave_search, query_ddg_search, query_google_patents, query_google_scholar
|
| 18 |
from ops import OPSBulkResponse, OPSNotConfigured, ops_scrap_patent, ops_scrap_patent_bulk, ops_search, token_manager as ops_token_manager
|
| 19 |
from utils import log_gathered_exceptions
|
| 20 |
from mcp_server import mount_mcp_server
|
| 21 |
|
| 22 |
+
# Anchored version of serp.py's PATENT_ID_CORE: validates a whole
|
| 23 |
+
# user-supplied patent id rather than finding one inside free text. Rejects
|
| 24 |
+
# obviously-malformed input (e.g. containing "/" or "?") with a clean 422
|
| 25 |
+
# instead of it reaching Google Patents/OPS as a confusing request.
|
| 26 |
+
PatentId = Annotated[str, StringConstraints(pattern=rf"^{PATENT_ID_CORE}$")]
|
| 27 |
+
|
| 28 |
logging.basicConfig(
|
| 29 |
level=logging.INFO,
|
| 30 |
format='[%(asctime)s][%(levelname)s][%(filename)s:%(lineno)d]: %(message)s',
|
|
|
|
| 241 |
|
| 242 |
|
| 243 |
@scrap_router.get("/scrap_patent/{patent_id}")
|
| 244 |
+
async def scrap_patent(patent_id: PatentId) -> PatentScrapResult:
|
| 245 |
"""Scraps the specified patent from Google Patents.
|
| 246 |
|
| 247 |
Falls back to the EPO OPS API (which covers patents missing from Google
|
|
|
|
| 280 |
|
| 281 |
class ScrapPatentsRequest(BaseModel):
|
| 282 |
"""Request model for scrapping multiple patents."""
|
| 283 |
+
patent_ids: list[PatentId] = Field(...,
|
| 284 |
+
description="List of patent IDs to scrap")
|
| 285 |
|
| 286 |
|
| 287 |
@scrap_router.post("/scrap_patents_bulk", response_model=PatentScrapBulkResponse)
|
|
|
|
| 302 |
|
| 303 |
|
| 304 |
@ops_router.get("/scrap_patent/{patent_id}")
|
| 305 |
+
async def ops_get_patent(patent_id: PatentId) -> PatentScrapResult:
|
| 306 |
"""Retrieves a patent (biblio + abstract + claims + description) via EPO OPS."""
|
| 307 |
if not ops_token_manager.configured:
|
| 308 |
raise HTTPException(
|
|
@@ -124,8 +124,14 @@ async def query_google_scholar(browser: Browser, q: str, n_results: int = 10):
|
|
| 124 |
return await _extract_google_scholar_results(page, n_results)
|
| 125 |
|
| 126 |
|
| 127 |
-
#
|
| 128 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 129 |
|
| 130 |
|
| 131 |
async def _extract_google_patents_results(page: Page, n_results: int) -> list[dict]:
|
|
|
|
| 124 |
return await _extract_google_scholar_results(page, n_results)
|
| 125 |
|
| 126 |
|
| 127 |
+
# A patent id's shape, e.g. "US11930446B2" or "EP4760514A1": a two-letter
|
| 128 |
+
# country code, at least 6 digits, and an optional kind-code suffix (a
|
| 129 |
+
# letter plus 0-2 digits). PATENT_ID_REGEX finds one inside free text (used
|
| 130 |
+
# to pull ids out of scraped search results); app.py reuses the same core
|
| 131 |
+
# pattern, anchored, to validate a user-supplied patent id - one definition
|
| 132 |
+
# so the two can't drift apart.
|
| 133 |
+
PATENT_ID_CORE = r"[A-Z]{2}\d{6,}(?:[A-Z]\d?)?"
|
| 134 |
+
PATENT_ID_REGEX = rf"\b{PATENT_ID_CORE}\b"
|
| 135 |
|
| 136 |
|
| 137 |
async def _extract_google_patents_results(page: Page, n_results: int) -> list[dict]:
|
|
@@ -14,6 +14,7 @@ import pytest
|
|
| 14 |
from httpx import ASGITransport
|
| 15 |
|
| 16 |
import app as app_module
|
|
|
|
| 17 |
from serp import SerpQuery
|
| 18 |
|
| 19 |
|
|
@@ -146,6 +147,42 @@ async def test_search_patents_no_fallback_when_ops_not_configured(client, monkey
|
|
| 146 |
assert data["error"] is None
|
| 147 |
|
| 148 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 149 |
async def test_ops_keyword_search_flattens_results(client, monkeypatch):
|
| 150 |
async def fake_ops_search(client_arg, q, n):
|
| 151 |
return [{"title": f"OPS {q}", "body": "", "href": "", "id": "1"}]
|
|
|
|
| 14 |
from httpx import ASGITransport
|
| 15 |
|
| 16 |
import app as app_module
|
| 17 |
+
from scrap import PatentScrapResult
|
| 18 |
from serp import SerpQuery
|
| 19 |
|
| 20 |
|
|
|
|
| 147 |
assert data["error"] is None
|
| 148 |
|
| 149 |
|
| 150 |
+
# ------------------------------- patent_id validation -------------------------------
|
| 151 |
+
|
| 152 |
+
|
| 153 |
+
async def test_scrap_patent_rejects_malformed_patent_id(client):
|
| 154 |
+
resp = await client.get("/scrap/scrap_patent/not-a-patent-id")
|
| 155 |
+
assert resp.status_code == 422
|
| 156 |
+
|
| 157 |
+
|
| 158 |
+
async def test_scrap_patent_accepts_well_formed_patent_id(client, monkeypatch):
|
| 159 |
+
async def fake_scrap(client_arg, url):
|
| 160 |
+
return PatentScrapResult(title="Widget apparatus")
|
| 161 |
+
|
| 162 |
+
monkeypatch.setattr(app_module, "scrap_patent_async", fake_scrap)
|
| 163 |
+
|
| 164 |
+
resp = await client.get("/scrap/scrap_patent/US11930446B2")
|
| 165 |
+
|
| 166 |
+
assert resp.status_code == 200
|
| 167 |
+
assert resp.json()["title"] == "Widget apparatus"
|
| 168 |
+
|
| 169 |
+
|
| 170 |
+
async def test_ops_get_patent_rejects_malformed_patent_id(client):
|
| 171 |
+
resp = await client.get("/ops/scrap_patent/not-a-patent-id")
|
| 172 |
+
assert resp.status_code == 422
|
| 173 |
+
|
| 174 |
+
|
| 175 |
+
async def test_scrap_patents_bulk_rejects_malformed_id_in_list(client):
|
| 176 |
+
resp = await client.post(
|
| 177 |
+
"/scrap/scrap_patents_bulk", json={"patent_ids": ["US11930446B2", "garbage!!"]})
|
| 178 |
+
assert resp.status_code == 422
|
| 179 |
+
|
| 180 |
+
|
| 181 |
+
async def test_ops_scrap_patents_bulk_rejects_malformed_id_in_list(client):
|
| 182 |
+
resp = await client.post("/ops/scrap_patents_bulk", json={"patent_ids": ["garbage!!"]})
|
| 183 |
+
assert resp.status_code == 422
|
| 184 |
+
|
| 185 |
+
|
| 186 |
async def test_ops_keyword_search_flattens_results(client, monkeypatch):
|
| 187 |
async def fake_ops_search(client_arg, q, n):
|
| 188 |
return [{"title": f"OPS {q}", "body": "", "href": "", "id": "1"}]
|