oc-lode 2.1.0__py3-none-any.whl
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- lode/__init__.py +2 -0
- lode/api.py +553 -0
- lode/builder.py +307 -0
- lode/cli.py +99 -0
- lode/exceptions.py +43 -0
- lode/helpers/__init__.py +1 -0
- lode/helpers/formats.py +32 -0
- lode/helpers/html_utils.py +32 -0
- lode/helpers/rdf.py +89 -0
- lode/helpers/spool.py +86 -0
- lode/models/__init__.py +38 -0
- lode/models/annotation.py +9 -0
- lode/models/atom.py +28 -0
- lode/models/attribute.py +13 -0
- lode/models/cardinality.py +29 -0
- lode/models/collection.py +29 -0
- lode/models/concept.py +156 -0
- lode/models/container.py +23 -0
- lode/models/datatype.py +9 -0
- lode/models/datatypeRestriction.py +28 -0
- lode/models/individual.py +40 -0
- lode/models/literal.py +40 -0
- lode/models/model.py +73 -0
- lode/models/oneOf.py +20 -0
- lode/models/property.py +96 -0
- lode/models/propertyConceptRestriction.py +37 -0
- lode/models/propertySelfRestriction.py +20 -0
- lode/models/quantifier.py +22 -0
- lode/models/relation.py +104 -0
- lode/models/resource.py +224 -0
- lode/models/restriction.py +10 -0
- lode/models/rule.py +27 -0
- lode/models/statement.py +49 -0
- lode/models/truthFunction.py +32 -0
- lode/models/value.py +28 -0
- lode/models/variable.py +9 -0
- lode/reader/README.md +215 -0
- lode/reader/__init__.py +16 -0
- lode/reader/config/base.yaml +201 -0
- lode/reader/config/owl.yaml +598 -0
- lode/reader/config/rdf.yaml +72 -0
- lode/reader/config/rdfs.yaml +69 -0
- lode/reader/config/skos.yaml +22 -0
- lode/reader/config_manager.py +295 -0
- lode/reader/loader.py +278 -0
- lode/reader/logic/__init__.py +15 -0
- lode/reader/logic/base_logic.py +892 -0
- lode/reader/logic/owl_logic.py +1104 -0
- lode/reader/logic/rdf_logic.py +253 -0
- lode/reader/logic/skos_logic.py +91 -0
- lode/reader/modules.py +229 -0
- lode/reader/provenance_graph.py +60 -0
- lode/reader/reader.py +292 -0
- lode/reader/security.py +221 -0
- lode/reader/warnings/__init__.py +15 -0
- lode/reader/warnings/owl_warnings.py +171 -0
- lode/static/style.css +1087 -0
- lode/templates/_card_owl.html +307 -0
- lode/templates/_card_skos.html +50 -0
- lode/templates/_toc_sidebar.html +95 -0
- lode/templates/_value.html +55 -0
- lode/templates/error.html +74 -0
- lode/templates/index.html +499 -0
- lode/templates/viewer.html +500 -0
- lode/viewer/__init__.py +28 -0
- lode/viewer/base_viewer.py +775 -0
- lode/viewer/config/model_annotations_deterministic_sorting.yaml +42 -0
- lode/viewer/config/owl_viewer.yaml +49 -0
- lode/viewer/config/rdf_viewer.yaml +39 -0
- lode/viewer/config/skos_viewer.yaml +17 -0
- lode/viewer/helpers.py +190 -0
- lode/viewer/owl_viewer.py +187 -0
- lode/viewer/rdf_viewer.py +8 -0
- lode/viewer/skos_viewer.py +172 -0
- oc_lode-2.1.0.dist-info/METADATA +169 -0
- oc_lode-2.1.0.dist-info/RECORD +79 -0
- oc_lode-2.1.0.dist-info/WHEEL +4 -0
- oc_lode-2.1.0.dist-info/entry_points.txt +2 -0
- oc_lode-2.1.0.dist-info/licenses/LICENSE.md +15 -0
lode/__init__.py
ADDED
lode/api.py
ADDED
|
@@ -0,0 +1,553 @@
|
|
|
1
|
+
# lode/api.py
|
|
2
|
+
"""LODE 2.0 — HTTP API.
|
|
3
|
+
|
|
4
|
+
Endpoints:
|
|
5
|
+
GET / web form
|
|
6
|
+
GET /health liveness probe
|
|
7
|
+
GET /extract render an artefact (HTML, or RDF via content negotiation)
|
|
8
|
+
POST /extract same, from an uploaded file
|
|
9
|
+
GET /build static documentation site, as a ZIP archive
|
|
10
|
+
POST /build same, from an uploaded file
|
|
11
|
+
"""
|
|
12
|
+
|
|
13
|
+
# ==========================================================================
|
|
14
|
+
# 1. IMPORTS
|
|
15
|
+
# ==========================================================================
|
|
16
|
+
from pathlib import Path
|
|
17
|
+
from typing import Optional
|
|
18
|
+
from urllib.parse import urlencode
|
|
19
|
+
import hashlib
|
|
20
|
+
import logging
|
|
21
|
+
import os
|
|
22
|
+
import re
|
|
23
|
+
import shutil
|
|
24
|
+
import tempfile
|
|
25
|
+
import threading
|
|
26
|
+
import time
|
|
27
|
+
import traceback
|
|
28
|
+
from pathlib import Path
|
|
29
|
+
|
|
30
|
+
from fastapi import FastAPI, File, Form, HTTPException, Query, Request, UploadFile
|
|
31
|
+
from fastapi.responses import FileResponse, HTMLResponse, Response
|
|
32
|
+
from fastapi.staticfiles import StaticFiles
|
|
33
|
+
from fastapi.templating import Jinja2Templates
|
|
34
|
+
from starlette.background import BackgroundTask
|
|
35
|
+
from starlette.concurrency import run_in_threadpool
|
|
36
|
+
from starlette.middleware.gzip import GZipMiddleware
|
|
37
|
+
from uvicorn.middleware.proxy_headers import ProxyHeadersMiddleware
|
|
38
|
+
|
|
39
|
+
# Internal modules
|
|
40
|
+
from lode.helpers.formats import (
|
|
41
|
+
ENABLED_FORMATS, ReadAsFormat,
|
|
42
|
+
check_format_enabled as _check_format_enabled,
|
|
43
|
+
)
|
|
44
|
+
from lode.helpers.html_utils import minify as _minify
|
|
45
|
+
import lode.helpers.spool as spool
|
|
46
|
+
from lode.reader import Reader
|
|
47
|
+
from lode.reader import security
|
|
48
|
+
from lode.exceptions import LODEError, ArtefactValidationError
|
|
49
|
+
from lode.viewer.base_viewer import SERIALIZATION_FORMATS
|
|
50
|
+
from lode.reader.provenance_graph import ProvenanceGraph
|
|
51
|
+
from lode.helpers.rdf import iri_local_name
|
|
52
|
+
|
|
53
|
+
|
|
54
|
+
# ==========================================================================
|
|
55
|
+
# 2. APP SETUP
|
|
56
|
+
# ==========================================================================
|
|
57
|
+
LOG_LEVEL = os.getenv("LODE_LOG_LEVEL", "WARNING").strip().upper()
|
|
58
|
+
|
|
59
|
+
logging.basicConfig(
|
|
60
|
+
level=LOG_LEVEL,
|
|
61
|
+
format="%(asctime)s - %(name)s - %(levelname)s - %(message)s",
|
|
62
|
+
)
|
|
63
|
+
# Third-party loggers: silence per-request HTTP noise and URI warnings.
|
|
64
|
+
logging.getLogger("urllib3").setLevel(logging.WARNING)
|
|
65
|
+
logging.getLogger("rdflib").setLevel(logging.ERROR)
|
|
66
|
+
logger = logging.getLogger(__name__)
|
|
67
|
+
|
|
68
|
+
# When enabled, error pages include the full traceback (development only).
|
|
69
|
+
DEBUG = os.getenv("LODE_DEBUG", "").strip().lower() in ("1", "true", "yes", "on")
|
|
70
|
+
|
|
71
|
+
TAGS_METADATA = [
|
|
72
|
+
{"name": "extract", "description": "Render a semantic artefact as HTML or serialized RDF."},
|
|
73
|
+
{"name": "build", "description": "Generate a full static documentation site, delivered as a ZIP."},
|
|
74
|
+
{"name": "ops", "description": "Operational endpoints: web form and health probe."},
|
|
75
|
+
]
|
|
76
|
+
|
|
77
|
+
app = FastAPI(
|
|
78
|
+
title="LODE 2.0 API",
|
|
79
|
+
version="0.3.12",
|
|
80
|
+
description="Live OWL Documentation Environment — extract and document "
|
|
81
|
+
"RDF/OWL/SKOS semantic artefacts as browsable HTML.",
|
|
82
|
+
license_info={"name": "ISC"},
|
|
83
|
+
contact={"name": "OpenCitations", "url": "https://github.com/opencitations/lode"},
|
|
84
|
+
openapi_tags=TAGS_METADATA,
|
|
85
|
+
)
|
|
86
|
+
|
|
87
|
+
# Compress HTML responses on the wire.
|
|
88
|
+
app.add_middleware(GZipMiddleware, minimum_size=1000)
|
|
89
|
+
# Fix "blocked loading mixed active content" on style.css behind a proxy.
|
|
90
|
+
app.add_middleware(ProxyHeadersMiddleware, trusted_hosts="*")
|
|
91
|
+
|
|
92
|
+
_PKG_DIR = Path(__file__).resolve().parent
|
|
93
|
+
templates = Jinja2Templates(directory=str(_PKG_DIR / "templates"))
|
|
94
|
+
app.mount("/static", StaticFiles(directory=str(_PKG_DIR / "static")), name="static")
|
|
95
|
+
|
|
96
|
+
# Beautify produced HTML.
|
|
97
|
+
templates.env.trim_blocks = True
|
|
98
|
+
templates.env.lstrip_blocks = True
|
|
99
|
+
|
|
100
|
+
# ==========================================================================
|
|
101
|
+
# 5. BUILD WORKSPACE — quota and concurrency for /build
|
|
102
|
+
# ==========================================================================
|
|
103
|
+
# Builds write hundreds of files each, so they get their own directory with an
|
|
104
|
+
# explicit budget instead of filling up /tmp.
|
|
105
|
+
BUILD_DIR = os.path.realpath(
|
|
106
|
+
os.getenv("LODE_BUILD_DIR", os.path.join(tempfile.gettempdir(), "lode-builds"))
|
|
107
|
+
)
|
|
108
|
+
os.makedirs(BUILD_DIR, exist_ok=True)
|
|
109
|
+
_BUILD_TTL = 30 * 60 # orphans (aborted downloads) age out
|
|
110
|
+
_BUILD_MAX_BYTES = 2 * 1024 ** 3 # 2 GB for all pending builds
|
|
111
|
+
_BUILD_SLOTS = threading.BoundedSemaphore(2) # concurrent builds, per worker process
|
|
112
|
+
_BUILD_WAIT = 5 # seconds a request waits for a slot
|
|
113
|
+
|
|
114
|
+
|
|
115
|
+
def _prune_builds() -> int:
|
|
116
|
+
"""Drop orphaned build directories and return the bytes still in use.
|
|
117
|
+
|
|
118
|
+
A build is orphaned when its BackgroundTask never fired, i.e. the client
|
|
119
|
+
opened the connection and never finished reading the archive.
|
|
120
|
+
"""
|
|
121
|
+
cutoff = time.time() - _BUILD_TTL
|
|
122
|
+
total = 0
|
|
123
|
+
for name in os.listdir(BUILD_DIR):
|
|
124
|
+
p = os.path.join(BUILD_DIR, name)
|
|
125
|
+
try:
|
|
126
|
+
st = os.stat(p)
|
|
127
|
+
except OSError:
|
|
128
|
+
continue
|
|
129
|
+
if st.st_mtime < cutoff:
|
|
130
|
+
shutil.rmtree(p, ignore_errors=True)
|
|
131
|
+
continue
|
|
132
|
+
for root, _, files in os.walk(p):
|
|
133
|
+
for f in files:
|
|
134
|
+
try:
|
|
135
|
+
total += os.path.getsize(os.path.join(root, f))
|
|
136
|
+
except OSError:
|
|
137
|
+
pass
|
|
138
|
+
return total
|
|
139
|
+
|
|
140
|
+
|
|
141
|
+
# ==========================================================================
|
|
142
|
+
# 6. HTML HELPERS
|
|
143
|
+
# ==========================================================================
|
|
144
|
+
|
|
145
|
+
def _nav_qs(read_as: str, url, upload_id, lang, imported, closure, partial_import) -> str:
|
|
146
|
+
"""Build the query string that keeps the current view parameters in navigation links.
|
|
147
|
+
|
|
148
|
+
Exactly one of `upload_id` and `url` is emitted, so a link never mixes the two
|
|
149
|
+
sources. Boolean options are included only when enabled.
|
|
150
|
+
"""
|
|
151
|
+
p = {"read_as": read_as, "lang": lang or ""}
|
|
152
|
+
p["upload_id" if upload_id else "url"] = upload_id or (url or "")
|
|
153
|
+
if imported:
|
|
154
|
+
p["imported"] = "true"
|
|
155
|
+
if closure:
|
|
156
|
+
p["closure"] = "true"
|
|
157
|
+
if partial_import:
|
|
158
|
+
p["partial_import"] = "true"
|
|
159
|
+
|
|
160
|
+
return urlencode(p)
|
|
161
|
+
|
|
162
|
+
def _render_view(request, reader, *, resource, lang, source_url, upload_id, read_as,
|
|
163
|
+
imported=None, closure=None, partial_import=None):
|
|
164
|
+
"""Render one resource (or the whole artefact) as the minified viewer page.
|
|
165
|
+
|
|
166
|
+
The body is minified after rendering, so Content-Length is recomputed:
|
|
167
|
+
keeping the old value would truncate the response.
|
|
168
|
+
"""
|
|
169
|
+
viewer = reader.get_viewer()
|
|
170
|
+
data = viewer.get_view_data(resource_uri=resource, language=lang)
|
|
171
|
+
data["warnings"] = viewer.format_warnings(reader.get_warnings())
|
|
172
|
+
resp = templates.TemplateResponse(request, "viewer.html", {
|
|
173
|
+
"source_url": source_url,
|
|
174
|
+
"upload_id": upload_id,
|
|
175
|
+
"nav_qs": _nav_qs(read_as, source_url, upload_id, lang, imported, closure, partial_import),
|
|
176
|
+
**data,
|
|
177
|
+
})
|
|
178
|
+
resp.body = _minify(resp.body.decode("utf-8")).encode("utf-8")
|
|
179
|
+
resp.headers["content-length"] = str(len(resp.body))
|
|
180
|
+
return resp
|
|
181
|
+
|
|
182
|
+
|
|
183
|
+
# ==========================================================================
|
|
184
|
+
# 7. READER RESOLUTION
|
|
185
|
+
# ==========================================================================
|
|
186
|
+
def _url_token(url, read_as, imported, closure, partial_import) -> str:
|
|
187
|
+
"""Return the stable cache key for a URL and its load options.
|
|
188
|
+
|
|
189
|
+
Every option that changes the loaded graph is part of the key, so two
|
|
190
|
+
requests share a cached copy only if they would produce the same result.
|
|
191
|
+
"""
|
|
192
|
+
key =f"{url}|{read_as}|{imported}|{closure}|{partial_import}".encode()
|
|
193
|
+
return "url_" + hashlib.sha256(key).hexdigest()[:32]
|
|
194
|
+
|
|
195
|
+
|
|
196
|
+
def _load_url(url, read_as, imported, closure, partial_import, warnings, use_cache=True):
|
|
197
|
+
"""Load an artefact from a URL, through the disk cache when allowed.
|
|
198
|
+
|
|
199
|
+
The URL is validated first (SSRF protection). Only the main document is
|
|
200
|
+
cached: modules (imports, closure) are re-applied on every hit, so the cache
|
|
201
|
+
never goes stale on remote ontologies. Blocking: network and disk I/O.
|
|
202
|
+
"""
|
|
203
|
+
# Enforce http(s)://host up front: a non-URL value (local path, file://, ...)
|
|
204
|
+
# must never reach the loader and be opened as a local file.
|
|
205
|
+
security.check_url_safe(url)
|
|
206
|
+
spool.prune()
|
|
207
|
+
token = _url_token(url, read_as, imported, closure, partial_import)
|
|
208
|
+
path = spool.get_path(token)
|
|
209
|
+
if use_cache and os.path.exists(path):
|
|
210
|
+
# cache hit: rebuild from the saved Turtle
|
|
211
|
+
reader = Reader()
|
|
212
|
+
reader.load_instances(path, read_as, imported=imported, closure=closure, partial_import=partial_import,
|
|
213
|
+
warnings=warnings, source_name=url)
|
|
214
|
+
return reader
|
|
215
|
+
if not use_cache:
|
|
216
|
+
# cache=false: drop the stale copy so the fresh fetch replaces it
|
|
217
|
+
try:
|
|
218
|
+
os.unlink(path)
|
|
219
|
+
except OSError:
|
|
220
|
+
pass
|
|
221
|
+
# cache miss (or forced refresh): fetch and process the URL
|
|
222
|
+
reader = Reader()
|
|
223
|
+
reader.load_instances(url, read_as, imported=imported, closure=closure, partial_import=partial_import,
|
|
224
|
+
warnings=warnings)
|
|
225
|
+
# persist ONLY the main document: modules are re-applied on every hit.
|
|
226
|
+
try:
|
|
227
|
+
g = reader._graph
|
|
228
|
+
main = g.graph(reader._ontology_iri) if isinstance(g, ProvenanceGraph) else g
|
|
229
|
+
with open(path, "wb") as f:
|
|
230
|
+
f.write(main.serialize(format="turtle").encode("utf-8"))
|
|
231
|
+
except OSError:
|
|
232
|
+
pass
|
|
233
|
+
return reader
|
|
234
|
+
|
|
235
|
+
def _resolve_reader(read_as: str, url, upload_id, imported, closure, partial_import, warnings,
|
|
236
|
+
use_cache=True):
|
|
237
|
+
"""Build a Reader from an uploaded file or from a URL.
|
|
238
|
+
|
|
239
|
+
Raises ArtefactValidationError when the upload has expired or when neither
|
|
240
|
+
source is given. Uploads win over URLs and are never re-fetched.
|
|
241
|
+
"""
|
|
242
|
+
if upload_id:
|
|
243
|
+
# Uploads are not re-fetched, so the cache flag does not apply to them.
|
|
244
|
+
path = spool.get_path(upload_id)
|
|
245
|
+
if not os.path.exists(path):
|
|
246
|
+
raise ArtefactValidationError("Upload expired, please re-upload",
|
|
247
|
+
context={"upload_id": upload_id})
|
|
248
|
+
reader = Reader()
|
|
249
|
+
reader.load_instances(path, read_as, imported=imported, closure=closure, partial_import=partial_import,
|
|
250
|
+
warnings=warnings)
|
|
251
|
+
return reader
|
|
252
|
+
if url:
|
|
253
|
+
return _load_url(url, read_as, imported, closure, partial_import, warnings, use_cache=use_cache)
|
|
254
|
+
raise ArtefactValidationError("Missing 'url' or 'upload_id'")
|
|
255
|
+
|
|
256
|
+
|
|
257
|
+
async def _read_validated_upload(file: UploadFile) -> bytes:
|
|
258
|
+
"""Run every upload check in one place, then return the raw bytes."""
|
|
259
|
+
security.check_extension(file.filename)
|
|
260
|
+
content = await security.read_upload_capped(file)
|
|
261
|
+
security.check_is_text(content)
|
|
262
|
+
security.check_safe_xml(content.decode("utf-8-sig"))
|
|
263
|
+
return content
|
|
264
|
+
|
|
265
|
+
|
|
266
|
+
# ==========================================================================
|
|
267
|
+
# 8. ZIP BUILD HELPERS
|
|
268
|
+
# ==========================================================================
|
|
269
|
+
def _slug(name: Optional[str]) -> str:
|
|
270
|
+
"""Safe directory/file stem derived from a URL or filename.
|
|
271
|
+
|
|
272
|
+
Must never yield '.' or '..': those would let make_archive(base_dir=...)
|
|
273
|
+
escape the build directory and archive its parent.
|
|
274
|
+
"""
|
|
275
|
+
if not name:
|
|
276
|
+
return "ontology"
|
|
277
|
+
raw = name.rstrip("/").split("#")[0].split("/")[-1]
|
|
278
|
+
raw = re.sub(r"\.(ttl|rdf|owl|n3|nt|jsonld|xml)$", "", raw, flags=re.I)
|
|
279
|
+
raw = re.sub(r"[^A-Za-z0-9._-]", "-", raw).strip("-.")
|
|
280
|
+
if not raw or raw in {".", ".."}:
|
|
281
|
+
return "ontology"
|
|
282
|
+
return raw[:64]
|
|
283
|
+
|
|
284
|
+
|
|
285
|
+
def _zip_response(reader, slug: str, lang: Optional[str] = None) -> FileResponse:
|
|
286
|
+
"""Render the whole static site and return it as a ZIP download.
|
|
287
|
+
|
|
288
|
+
Blocking and disk-heavy: callers must keep it off the event loop.
|
|
289
|
+
"""
|
|
290
|
+
from lode.builder import build_html # lazy: avoids a circular import
|
|
291
|
+
|
|
292
|
+
if not _BUILD_SLOTS.acquire(timeout=_BUILD_WAIT):
|
|
293
|
+
raise HTTPException(503, "Server busy: too many concurrent builds",
|
|
294
|
+
headers={"Retry-After": "30"})
|
|
295
|
+
workdir = None
|
|
296
|
+
try:
|
|
297
|
+
if _prune_builds() > _BUILD_MAX_BYTES:
|
|
298
|
+
raise HTTPException(503, "Build storage full, retry shortly",
|
|
299
|
+
headers={"Retry-After": "60"})
|
|
300
|
+
|
|
301
|
+
workdir = tempfile.mkdtemp(prefix="build_", dir=BUILD_DIR)
|
|
302
|
+
site_dir = Path(workdir) / slug
|
|
303
|
+
site_dir.mkdir(parents=True, exist_ok=True)
|
|
304
|
+
|
|
305
|
+
build_html(reader.get_viewer(), site_dir, lang=lang or "en", reader=reader)
|
|
306
|
+
|
|
307
|
+
zip_path = shutil.make_archive(
|
|
308
|
+
os.path.join(workdir, f"{slug}-lode"), "zip",
|
|
309
|
+
root_dir=workdir, base_dir=slug,
|
|
310
|
+
)
|
|
311
|
+
# The loose pages are redundant once archived: halve the disk peak.
|
|
312
|
+
shutil.rmtree(site_dir, ignore_errors=True)
|
|
313
|
+
except BaseException:
|
|
314
|
+
if workdir:
|
|
315
|
+
shutil.rmtree(workdir, ignore_errors=True)
|
|
316
|
+
raise
|
|
317
|
+
finally:
|
|
318
|
+
# Free the slot as soon as the build ends, not when the client is done
|
|
319
|
+
# downloading: a slow reader must not hold a build slot hostage.
|
|
320
|
+
_BUILD_SLOTS.release()
|
|
321
|
+
|
|
322
|
+
return FileResponse(
|
|
323
|
+
zip_path,
|
|
324
|
+
media_type="application/zip",
|
|
325
|
+
filename=f"{slug}-lode.zip",
|
|
326
|
+
background=BackgroundTask(shutil.rmtree, workdir, True),
|
|
327
|
+
)
|
|
328
|
+
|
|
329
|
+
|
|
330
|
+
# ==========================================================================
|
|
331
|
+
# 9. ERROR RENDERING
|
|
332
|
+
# ==========================================================================
|
|
333
|
+
def _error_payload(exc: Exception) -> dict:
|
|
334
|
+
"""Build the dict used by error.html for any exception.
|
|
335
|
+
|
|
336
|
+
The traceback is included only for non-LODE errors and only when LODE_DEBUG
|
|
337
|
+
is set: it is always logged server-side and never shown to users otherwise.
|
|
338
|
+
"""
|
|
339
|
+
is_lode = isinstance(exc, LODEError)
|
|
340
|
+
return {
|
|
341
|
+
"type": type(exc).__name__,
|
|
342
|
+
"message": str(exc),
|
|
343
|
+
"context": exc.context if is_lode else {},
|
|
344
|
+
"request_id": exc.request_id if is_lode else None,
|
|
345
|
+
# Never leak the traceback to the user in production: it is always logged
|
|
346
|
+
# server-side; it is only surfaced in the page when LODE_DEBUG is set.
|
|
347
|
+
"traceback": traceback.format_exc() if (not is_lode and DEBUG) else None,
|
|
348
|
+
}
|
|
349
|
+
|
|
350
|
+
|
|
351
|
+
@app.exception_handler(LODEError)
|
|
352
|
+
async def lode_error_handler(request: Request, exc: LODEError):
|
|
353
|
+
"""Show a domain error (LODEError) as an HTML page with HTTP 400."""
|
|
354
|
+
logger.error(f"{type(exc).__name__}: {exc}")
|
|
355
|
+
return templates.TemplateResponse(
|
|
356
|
+
"error.html", {"request": request, "error": _error_payload(exc)}, status_code=400
|
|
357
|
+
)
|
|
358
|
+
|
|
359
|
+
|
|
360
|
+
@app.exception_handler(Exception)
|
|
361
|
+
async def unexpected_error_handler(request: Request, exc: Exception):
|
|
362
|
+
"""Show an unhandled exception as an HTML page with HTTP 500.
|
|
363
|
+
|
|
364
|
+
The full traceback goes to the server log; the page shows it only in debug mode.
|
|
365
|
+
"""
|
|
366
|
+
logger.exception("Unexpected error")
|
|
367
|
+
return templates.TemplateResponse(
|
|
368
|
+
"error.html", {"request": request, "error": _error_payload(exc)}, status_code=500
|
|
369
|
+
)
|
|
370
|
+
|
|
371
|
+
|
|
372
|
+
# ==========================================================================
|
|
373
|
+
# 10. MIDDLEWARE
|
|
374
|
+
# ==========================================================================
|
|
375
|
+
@app.middleware("http")
|
|
376
|
+
async def limit_upload_size(request: Request, call_next):
|
|
377
|
+
"""Reject oversized POST bodies early, before reading them, when the client
|
|
378
|
+
declares a Content-Length. (Chunked bodies omit it: those are still capped
|
|
379
|
+
in read_upload_capped; the hard limit belongs at the reverse proxy.)"""
|
|
380
|
+
if request.method == "POST":
|
|
381
|
+
declared = request.headers.get("content-length")
|
|
382
|
+
if declared and declared.isdigit() and int(declared) > security.MAX_BYTES:
|
|
383
|
+
exc = ArtefactValidationError(
|
|
384
|
+
"File too large",
|
|
385
|
+
context={"bytes": int(declared), "max": security.MAX_BYTES},
|
|
386
|
+
)
|
|
387
|
+
return templates.TemplateResponse(
|
|
388
|
+
"error.html", {"request": request, "error": _error_payload(exc)},
|
|
389
|
+
status_code=400,
|
|
390
|
+
)
|
|
391
|
+
return await call_next(request)
|
|
392
|
+
|
|
393
|
+
|
|
394
|
+
# ==========================================================================
|
|
395
|
+
# 11. ROUTES — pages
|
|
396
|
+
# ==========================================================================
|
|
397
|
+
@app.get("/", response_class=HTMLResponse, tags=["ops"], summary="Web input form")
|
|
398
|
+
async def input_web_interface(request: Request):
|
|
399
|
+
"""Serve the HTML input form, the entry point of the web interface."""
|
|
400
|
+
return templates.TemplateResponse(request, "index.html", {
|
|
401
|
+
"formats": [format.value for format in ReadAsFormat],
|
|
402
|
+
"enabled_formats": sorted(ENABLED_FORMATS),
|
|
403
|
+
})
|
|
404
|
+
|
|
405
|
+
|
|
406
|
+
@app.get("/health", tags=["ops"], summary="Liveness probe")
|
|
407
|
+
async def health_check():
|
|
408
|
+
"""Return 200 with a fixed body. Used by Docker/K8s readiness checks."""
|
|
409
|
+
return {"status": "ok"}
|
|
410
|
+
|
|
411
|
+
|
|
412
|
+
# ==========================================================================
|
|
413
|
+
# 12. ROUTES — /extract
|
|
414
|
+
# ==========================================================================
|
|
415
|
+
@app.get("/extract", tags=["extract"], summary="Render an artefact as HTML or RDF")
|
|
416
|
+
def extract_get(
|
|
417
|
+
request: Request,
|
|
418
|
+
read_as: ReadAsFormat = Query(description="Read the source as this artefact type"),
|
|
419
|
+
url: Optional[str] = Query(None, description="Public http(s) URL of the artefact"),
|
|
420
|
+
upload_id: Optional[str] = Query(None, description="Token of a previously uploaded file"),
|
|
421
|
+
resource: Optional[str] = Query(None, description="IRI of a single resource; omit for the whole ontology"),
|
|
422
|
+
lang: Optional[str] = Query(None, description="Language tag for labels, e.g. 'en', 'it'"),
|
|
423
|
+
imported: Optional[bool] = Query(None, description="Also load owl:imports (depth 1)"),
|
|
424
|
+
closure: Optional[bool] = Query(None, description="Load the full owl:imports closure"),
|
|
425
|
+
partial_import: Optional[bool] = Query(None, description="Additionally load only explicitly mentioned entities from other ontologies"),
|
|
426
|
+
format: Optional[str] = Query(None, description="Force a serialization (ttl, rdf, jsonld…) instead of HTML"),
|
|
427
|
+
warnings: bool = Query(False, description="Collect and show reader warnings"),
|
|
428
|
+
cache: bool = Query(True, description="Use the disk cache for fetched URLs"),
|
|
429
|
+
):
|
|
430
|
+
"""Render an artefact. Returns HTML by default, or RDF when a serialization
|
|
431
|
+
is requested via the `format` query or the `Accept` header."""
|
|
432
|
+
_check_format_enabled(read_as)
|
|
433
|
+
|
|
434
|
+
reader = _resolve_reader(read_as.value, url, upload_id, imported, closure, partial_import,
|
|
435
|
+
warnings, use_cache=cache)
|
|
436
|
+
|
|
437
|
+
# Content negotiation: triples of the loaded artefact only, quads are never negotiable
|
|
438
|
+
accept = request.headers.get("accept", "text/html")
|
|
439
|
+
available = SERIALIZATION_FORMATS
|
|
440
|
+
serial = None
|
|
441
|
+
|
|
442
|
+
if format:
|
|
443
|
+
fmt_l = format.lower()
|
|
444
|
+
serial = next((f for f in available if f["ext"] == fmt_l), None)
|
|
445
|
+
if serial is None:
|
|
446
|
+
raise HTTPException(406, f"Format '{format}' not available for this request")
|
|
447
|
+
else:
|
|
448
|
+
for part in accept.split(","):
|
|
449
|
+
mime = part.split(";", 1)[0].strip()
|
|
450
|
+
serial = next((f for f in available if f["mime"] == mime), None)
|
|
451
|
+
if serial:
|
|
452
|
+
break
|
|
453
|
+
|
|
454
|
+
if serial:
|
|
455
|
+
if resource:
|
|
456
|
+
serialized = reader.get_viewer().export_resource(resource, serial["fmt"])
|
|
457
|
+
filename = iri_local_name(resource) or "resource"
|
|
458
|
+
else:
|
|
459
|
+
serialized = reader.get_main_graph().serialize(format=serial["fmt"])
|
|
460
|
+
filename = (iri_local_name(url) if url else "") or "graph"
|
|
461
|
+
return Response(content=serialized, media_type=serial["mime"],
|
|
462
|
+
headers={"Content-Disposition": f'inline; filename="{filename}.{serial["ext"]}"'})
|
|
463
|
+
|
|
464
|
+
logger.info("=== REQUEST SUCCESS ===")
|
|
465
|
+
return _render_view(request, reader, resource=resource, lang=lang,
|
|
466
|
+
source_url=url, upload_id=upload_id, read_as=read_as.value,
|
|
467
|
+
imported=imported, closure=closure, partial_import=partial_import)
|
|
468
|
+
|
|
469
|
+
|
|
470
|
+
@app.post("/extract", response_class=HTMLResponse, tags=["extract"],
|
|
471
|
+
summary="Render an uploaded artefact as HTML")
|
|
472
|
+
async def extract_post(
|
|
473
|
+
request: Request,
|
|
474
|
+
read_as: ReadAsFormat = Form(..., description="Read the uploaded file as this artefact type"),
|
|
475
|
+
file: UploadFile = File(..., description="The artefact file (ttl, rdf, owl…)"),
|
|
476
|
+
resource: Optional[str] = Form(None, description="IRI of a single resource; omit for the whole ontology"),
|
|
477
|
+
lang: Optional[str] = Form(None, description="Language tag for labels, e.g. 'en', 'it'"),
|
|
478
|
+
imported: Optional[bool] = Form(None, description="Also load owl:imports (depth 1)"),
|
|
479
|
+
partial_import: Optional[bool] = Form(None, description="Additionally load only explicitly mentioned entities from other ontologies"),
|
|
480
|
+
closure: Optional[bool] = Form(None, description="Load the full owl:imports closure"),
|
|
481
|
+
warnings: bool = Form(False, description="Collect and show reader warnings"),
|
|
482
|
+
):
|
|
483
|
+
"""Render an artefact from an uploaded file instead of a URL."""
|
|
484
|
+
logger.info("=== FILE UPLOAD START ===")
|
|
485
|
+
logger.info(f"Filename: {file.filename}")
|
|
486
|
+
logger.info(f"Format: {read_as.value}")
|
|
487
|
+
|
|
488
|
+
_check_format_enabled(read_as)
|
|
489
|
+
content = await _read_validated_upload(file)
|
|
490
|
+
token = spool.save(content)
|
|
491
|
+
|
|
492
|
+
reader = Reader()
|
|
493
|
+
await run_in_threadpool(
|
|
494
|
+
reader.load_instances, spool.get_path(token), read_as.value,
|
|
495
|
+
imported=imported, closure=closure, partial_import=partial_import, warnings=warnings,
|
|
496
|
+
)
|
|
497
|
+
return _render_view(request, reader, resource=resource, lang=lang,
|
|
498
|
+
source_url=None, upload_id=token, read_as=read_as.value,
|
|
499
|
+
imported=imported, closure=closure, partial_import=partial_import)
|
|
500
|
+
|
|
501
|
+
|
|
502
|
+
# ==========================================================================
|
|
503
|
+
# 13. ROUTES — /build
|
|
504
|
+
# ==========================================================================
|
|
505
|
+
@app.get("/build", tags=["build"], summary="Build a static site (ZIP) from a URL")
|
|
506
|
+
def build_get(
|
|
507
|
+
read_as: ReadAsFormat = Query(description="Read the source as this artefact type"),
|
|
508
|
+
url: Optional[str] = Query(None, description="Public http(s) URL of the artefact"),
|
|
509
|
+
upload_id: Optional[str] = Query(None, description="Token of a previously uploaded file"),
|
|
510
|
+
lang: Optional[str] = Query(None, description="Language tag for labels, e.g. 'en', 'it'"),
|
|
511
|
+
imported: Optional[bool] = Query(None, description="Also load owl:imports (depth 1)"),
|
|
512
|
+
closure: Optional[bool] = Query(None, description="Load the full owl:imports closure"),
|
|
513
|
+
partial_import: Optional[bool] = Query(None, description="Additionally load only explicitly mentioned entities from other ontologies"),
|
|
514
|
+
warnings: bool = Query(False, description="Collect and show reader warnings"),
|
|
515
|
+
cache: bool = Query(True, description="Use the disk cache for fetched URLs"),
|
|
516
|
+
):
|
|
517
|
+
"""Static documentation site as a ZIP. Sync on purpose: FastAPI runs
|
|
518
|
+
non-async endpoints in the threadpool, keeping the event loop free."""
|
|
519
|
+
_check_format_enabled(read_as)
|
|
520
|
+
reader = _resolve_reader(read_as.value, url, upload_id, imported, closure, partial_import,
|
|
521
|
+
warnings, use_cache=cache)
|
|
522
|
+
return _zip_response(reader, _slug(url), lang)
|
|
523
|
+
|
|
524
|
+
|
|
525
|
+
@app.post("/build", tags=["build"], summary="Build a static site (ZIP) from an upload")
|
|
526
|
+
async def build_post(
|
|
527
|
+
read_as: ReadAsFormat = Form(..., description="Read the uploaded file as this artefact type"),
|
|
528
|
+
file: UploadFile = File(..., description="The artefact file (ttl, rdf, owl…)"),
|
|
529
|
+
lang: Optional[str] = Form(None, description="Language tag for labels, e.g. 'en', 'it'"),
|
|
530
|
+
imported: Optional[bool] = Form(None, description="Also load owl:imports (depth 1)"),
|
|
531
|
+
closure: Optional[bool] = Form(None, description="Load the full owl:imports closure"),
|
|
532
|
+
partial_import: Optional[bool] = Form(None, description="Additionally load only explicitly mentioned entities from other ontologies"),
|
|
533
|
+
warnings: bool = Form(False, description="Collect and show reader warnings"),
|
|
534
|
+
):
|
|
535
|
+
"""Build the static site from an uploaded file instead of a URL."""
|
|
536
|
+
_check_format_enabled(read_as)
|
|
537
|
+
content = await _read_validated_upload(file)
|
|
538
|
+
token = spool.save(content)
|
|
539
|
+
|
|
540
|
+
reader = Reader()
|
|
541
|
+
await run_in_threadpool(
|
|
542
|
+
reader.load_instances, spool.get_path(token), read_as.value,
|
|
543
|
+
imported=imported, closure=closure, partial_import=partial_import, warnings=warnings,
|
|
544
|
+
)
|
|
545
|
+
return await run_in_threadpool(_zip_response, reader, _slug(file.filename), lang)
|
|
546
|
+
|
|
547
|
+
|
|
548
|
+
# ==========================================================================
|
|
549
|
+
# 14. ENTRY POINT
|
|
550
|
+
# ==========================================================================
|
|
551
|
+
if __name__ == "__main__":
|
|
552
|
+
import uvicorn
|
|
553
|
+
uvicorn.run(app, host="0.0.0.0", port=8000)
|