1 # Fedi API Block - An aggregator for fetching blocking data from fediverse nodes
2 # Copyright (C) 2023 Free Software Foundation
4 # This program is free software: you can redistribute it and/or modify
5 # it under the terms of the GNU Affero General Public License as published
6 # by the Free Software Foundation, either version 3 of the License, or
7 # (at your option) any later version.
9 # This program is distributed in the hope that it will be useful,
10 # but WITHOUT ANY WARRANTY; without even the implied warranty of
11 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
12 # GNU Affero General Public License for more details.
14 # You should have received a copy of the GNU Affero General Public License
15 # along with this program. If not, see <https://www.gnu.org/licenses/>.
24 from fba import database
27 from fba.helpers import blacklist
28 from fba.helpers import cache
29 from fba.helpers import config
30 from fba.helpers import domain as domain_helper
32 from fba.http import federation
33 from fba.http import network
35 from fba.models import error_log
37 logging.basicConfig(level=logging.INFO)
38 logger = logging.getLogger(__name__)
40 # Found info from node, such as nodeinfo URL, detection mode that needs to be
41 # written to database. Both arrays must be filled at the same time or else
42 # update_data() will fail
45 # NULL means all detection methods have failed (maybe still reachable instance)
46 "detection_mode" : {},
51 # Last fetched instances
52 "last_instance_fetch": {},
57 # Last nodeinfo (fetched)
60 "last_status_code" : {},
62 "last_error_details" : {},
63 # Wether obfuscation has been used
64 "has_obfuscation" : {},
67 def _set_data(key: str, domain: str, value: any):
68 logger.debug("key='%s',domain='%s',value[]='%s' - CALLED!", key, domain, type(value))
69 domain_helper.raise_on(domain)
70 if not isinstance(key, str):
71 raise ValueError("Parameter key[]='{type(key)}' is not 'str'")
73 raise ValueError("Parameter 'key' is empty")
74 elif not key in _pending:
75 raise ValueError(f"key='{key}' not found in _pending")
76 elif not utils.is_primitive(value):
77 raise ValueError(f"value[]='{type(value)}' is not a primitive type")
80 _pending[key][domain] = value
84 def has_pending(domain: str) -> bool:
85 logger.debug("domain='%s' - CALLED!", domain)
86 domain_helper.raise_on(domain)
90 logger.debug("key='%s',domain='%s',_pending[key]()=%d", key, domain, len(_pending[key]))
91 if domain in _pending[key]:
95 logger.debug("has='%s' - EXIT!", has)
98 def update_data(domain: str):
99 logger.debug("domain='%s' - CALLED!", domain)
100 domain_helper.raise_on(domain)
101 if not has_pending(domain):
102 raise Exception(f"domain='{domain}' has no pending instance data, but function invoked")
103 elif not is_registered(domain):
104 raise Exception(f"domain='{domain}' cannot be updated while not being registered")
106 logger.debug("Updating instance data for domain='%s' ...", domain)
110 logger.debug("Checking key='%s',domain='%s'", key, domain)
111 if domain in _pending[key]:
112 logger.debug("Adding '%s' for key='%s' ...", _pending[key][domain], key)
113 fields.append(_pending[key][domain])
114 sql_string += f" {key} = ?,"
116 logger.debug("sql_string()=%d", len(sql_string))
118 raise ValueError(f"No fields have been set, but method invoked, domain='{domain}'")
120 # Set last_updated to current timestamp
121 fields.append(time.time())
123 # For WHERE statement
124 fields.append(domain)
126 logger.debug("sql_string='%s',fields()=%d", sql_string, len(fields))
127 sql_string = "UPDATE instances SET" + sql_string + " last_updated = ? WHERE domain = ? LIMIT 1"
129 logger.debug("Executing SQL: '%s'", sql_string)
130 database.cursor.execute(sql_string, fields)
132 logger.debug("rowcount=%d", database.cursor.rowcount)
133 if database.cursor.rowcount == 0:
134 raise Exception(f"Did not update any rows: domain='{domain}',fields()={len(fields)}")
136 logger.debug("Invoking commit() ...")
137 database.connection.commit()
139 logger.debug("Deleting _pending for domain='%s'", domain)
141 logger.debug("domain='%s',key='%s'", domain, key)
142 if domain in _pending[key]:
143 logger.debug("Deleting key='%s',domain='%s' ...", key, domain)
144 del _pending[key][domain]
146 logger.debug("EXIT!")
148 def add(domain: str, origin: str, command: str, path: str = None, software: str = None):
149 logger.debug("domain='%s',origin='%s',command='%s',path='%s',software='%s' - CALLED!", domain, origin, command, path, software)
150 domain_helper.raise_on(domain)
152 if not isinstance(origin, str) and origin is not None:
153 raise ValueError(f"origin[]='{type(origin)}' is not 'str'")
155 raise ValueError("Parameter 'origin' is empty")
156 elif not isinstance(command, str):
157 raise ValueError(f"command[]='{type(command)}' is not 'str'")
159 raise ValueError("Parameter 'command' is empty")
160 elif not isinstance(path, str) and path is not None:
161 raise ValueError(f"path[]='{type(path)}' is not 'str'")
163 raise ValueError("Parameter 'path' is empty")
164 elif not isinstance(software, str) and software is not None:
165 raise ValueError(f"software[]='{type(software)}' is not 'str'")
167 raise ValueError("Parameter 'software' is empty")
168 elif origin is not None and not validators.domain(origin.split("/")[0]):
169 raise ValueError(f"Bad origin name='{origin}'")
170 elif blacklist.is_blacklisted(domain):
171 raise Exception(f"domain='{domain}' is blacklisted, but method invoked")
172 elif domain.find("/profile/") > 0 or domain.find("/users/") > 0 or (is_registered(domain.split("/")[0]) and domain.find("/c/") > 0):
173 raise Exception(f"domain='{domain}' is a single user")
174 elif domain.find("/tag/") > 0:
175 raise Exception(f"domain='{domain}' is a tag")
179 logger.debug("domain='%s',origin='%s',command='%s',path='%s'", domain, origin, command, path)
180 software = federation.determine_software(domain, path)
181 except network.exceptions as exception:
182 logger.warning("Exception '%s' during determining software type, domain='%s'", type(exception), domain)
183 set_last_error(domain, exception)
185 logger.debug("Determined software='%s'", software)
186 if software == "lemmy" and domain.find("/c/") > 0:
187 domain = domain.split("/c/")[0]
188 if is_registered(domain):
189 logger.warning("domain='%s' already registered after cutting off user part. - EXIT!", domain)
192 logger.info("Adding instance domain='%s',origin='%s',software='%s',command='%s'", domain, origin, software, command)
193 database.cursor.execute(
194 "INSERT INTO instances (domain, origin, command, hash, software, first_seen) VALUES (?, ?, ?, ?, ?, ?)",
199 utils.get_hash(domain),
205 logger.debug("Marking domain='%s' as registered.", domain)
206 cache.set_sub_key("is_registered", domain, True)
208 logger.debug("Checking if domain='%s' has pending updates ...", domain)
209 if has_pending(domain):
210 logger.debug("Flushing updates for domain='%s' ...", domain)
213 logger.debug("EXIT!")
215 def set_last_nodeinfo(domain: str):
216 logger.debug("domain='%s' - CALLED!", domain)
217 domain_helper.raise_on(domain)
219 logger.debug("Updating last_nodeinfo for domain='%s'", domain)
220 _set_data("last_nodeinfo", domain, time.time())
222 logger.debug("EXIT!")
224 def set_last_error(domain: str, error: dict):
225 logger.debug("domain='%s',error[]='%s' - CALLED!", domain, type(error))
226 domain_helper.raise_on(domain)
228 logger.debug("error[]='%s' - BEFORE!", type(error))
229 if isinstance(error, (BaseException, json.decoder.JSONDecodeError)):
230 error = f"error[{type(error)}]='{str(error)}'"
231 logger.debug("error[]='%s' - AFTER!", type(error))
233 if isinstance(error, str):
234 logger.debug("Setting last_error_details='%s' (str)", error)
235 _set_data("last_status_code" , domain, 999)
236 _set_data("last_error_details", domain, error if error != "" else None)
237 elif isinstance(error, requests.models.Response):
238 logger.debug("Setting last_error_details='%s' (Response)", error.reason)
239 _set_data("last_status_code" , domain, error.status_code)
240 _set_data("last_error_details", domain, error.reason if error.reason != "" else None)
241 elif not isinstance(error, dict):
242 raise KeyError(f"Cannot handle keys in error[{type(error)}]='{error}'")
243 elif "status_code" in error and "error_message" in error:
244 logger.debug("Setting last_error_details='%s' (error_message)", error['error_message'])
245 _set_data("last_status_code" , domain, error["status_code"])
246 _set_data("last_error_details", domain, error["error_message"] if error["error_message"] != "" else None)
247 elif "json" in error and "error" in error["json"]:
248 logger.debug("Setting last_error_details='%s' (json,error)", error["json"]["error"])
249 _set_data("last_status_code" , domain, error["status_code"])
250 _set_data("last_error_details", domain, error["json"]["error"] if error["json"]["error"] != "" else None)
252 logger.debug("Invoking error_log.add(domain='%s',error[]='%s'", domain, type(error))
253 error_log.add(domain, error)
255 logger.debug("EXIT!")
257 def set_success(domain: str):
258 logger.debug("domain='%s' - CALLED!", domain)
259 domain_helper.raise_on(domain)
261 # Set both to success
262 _set_data("last_status_code" , domain, 200)
263 _set_data("last_error_details", domain, None)
265 logger.debug("EXIT!")
267 def is_registered(domain: str) -> bool:
268 logger.debug("domain='%s' - CALLED!", domain)
269 domain_helper.raise_on(domain)
271 logger.debug("domain='%s' - CALLED!", domain)
272 if not cache.key_exists("is_registered"):
273 logger.debug("Cache for 'is_registered' not initialized, fetching all rows ...")
274 database.cursor.execute("SELECT domain FROM instances")
277 cache.set_all("is_registered", database.cursor.fetchall(), True)
280 registered = cache.sub_key_exists("is_registered", domain)
282 logger.debug("registered='%s' - EXIT!", registered)
285 def is_recent(domain: str, column: str = "last_instance_fetch") -> bool:
286 logger.debug("domain='%s',column='%s' - CALLED!", domain)
287 domain_helper.raise_on(domain)
289 if not isinstance(column, str):
290 raise ValueError(f"Parameter column[]='{type(column)}' is not 'str'")
291 elif column not in ["last_instance_fetch", "last_blocked"]:
292 raise ValueError(f"Parameter column='{column}' is not expected")
293 elif not is_registered(domain):
294 logger.debug("domain='%s' is not registered, returning False - EXIT!", domain)
298 database.cursor.execute(f"SELECT {column} FROM instances WHERE domain = ? LIMIT 1", [domain])
301 fetched = database.cursor.fetchone()[0]
303 logger.debug("fetched[%s]='%s'", type(fetched), fetched)
304 recently = isinstance(fetched, float) and time.time() - fetched <= config.get("recheck_instance")
306 logger.debug("recently='%s' - EXIT!", recently)
309 def deobfuscate(char: str, domain: str, blocked_hash: str = None) -> tuple:
310 logger.debug("char='%s',domain='%s',blocked_hash='%s' - CALLED!", char, domain, blocked_hash)
312 if not isinstance(domain, str):
313 raise ValueError(f"Parameter domain[]='{type(domain)}' is not 'str'")
315 raise ValueError("Parameter 'domain' is empty")
316 elif domain.lower() != domain:
317 raise ValueError(f"Parameter domain='{domain}' must be all lower-case")
318 elif domain.endswith(".arpa"):
319 raise ValueError(f"domain='{domain}' is a domain for reversed IP addresses, please don't crawl them!")
320 elif domain.endswith(".tld"):
321 raise ValueError(f"domain='{domain}' is a fake domain, please don't crawl them!")
322 elif not isinstance(char, str):
323 raise ValueError(f"Parameter char[]='{type(char)}' is not 'str'")
325 raise ValueError("Parameter 'char' is empty")
326 elif not char in domain:
327 raise ValueError(f"char='{char}' not found in domain='{domain}' but function invoked")
328 elif not isinstance(blocked_hash, str) and blocked_hash is not None:
329 raise ValueError(f"Parameter blocked_hash[]='{type(blocked_hash)}' is not 'str'")
331 logger.debug("blocked_hash[]='%s'", type(blocked_hash))
332 if isinstance(blocked_hash, str):
333 logger.debug("Looking up blocked_hash='%s',domain='%s' ...", blocked_hash, domain)
334 database.cursor.execute(
335 "SELECT domain, origin, nodeinfo_url FROM instances WHERE hash = ? OR domain LIKE ? LIMIT 1", [blocked_hash, domain.replace(char, "_")]
338 row = database.cursor.fetchone()
339 logger.debug("row[]='%s'", type(row))
342 logger.debug("blocked_hash='%s' not found, trying domain='%s' ...", blocked_hash, domain)
343 return deobfuscate(char, domain)
345 logger.debug("Looking up domain='%s' ...", domain)
346 database.cursor.execute(
347 "SELECT domain, origin, nodeinfo_url FROM instances WHERE domain LIKE ? ORDER BY rowid LIMIT 1", [domain.replace(char, "_")]
350 row = database.cursor.fetchone()
351 logger.debug("row[]='%s'", type(row))
353 logger.debug("row[]='%s' - EXIT!", type(row))
356 def set_last_blocked(domain: str):
357 logger.debug("domain='%s' - CALLED!", domain)
358 domain_helper.raise_on(domain)
361 _set_data("last_blocked", domain, time.time())
362 logger.debug("EXIT!")
364 def set_last_instance_fetch(domain: str):
365 logger.debug("domain='%s' - CALLED!", domain)
366 domain_helper.raise_on(domain)
369 _set_data("last_instance_fetch", domain, time.time())
370 logger.debug("EXIT!")
372 def set_total_peers(domain: str, peers: list):
373 logger.debug("domain='%s',peers()=%d - CALLED!", domain, len(peers))
374 domain_helper.raise_on(domain)
376 if not isinstance(peers, list):
377 raise ValueError(f"Parameter peers[]='{type(peers)}' is not 'list': '%s'")
380 _set_data("total_peers", domain, len(peers))
381 logger.debug("EXIT!")
383 def set_nodeinfo_url(domain: str, url: str):
384 logger.debug("domain='%s',url='%s' - CALLED!", domain, url)
385 domain_helper.raise_on(domain)
387 if not isinstance(url, str):
388 raise ValueError("Parameter url[]='{type(url)}' is not 'list'")
390 raise ValueError("Parameter 'url' is empty")
393 _set_data("nodeinfo_url", domain, url)
394 logger.debug("EXIT!")
396 def set_detection_mode(domain: str, mode: str):
397 logger.debug("domain='%s',mode='%s' - CALLED!", domain, mode)
398 domain_helper.raise_on(domain)
400 if not isinstance(mode, str):
401 raise ValueError("Parameter mode[]='{type(mode)}' is not 'list'")
403 raise ValueError("Parameter 'mode' is empty")
406 _set_data("detection_mode", domain, mode)
407 logger.debug("EXIT!")
409 def set_has_obfuscation(domain: str, status: bool):
410 logger.debug("domain(%d)='%s',status='%s' - CALLED!", len(domain), domain, status)
411 domain_helper.raise_on(domain)
413 if not isinstance(status, bool):
414 raise ValueError(f"Parameter status[]='{type(status)}' is not 'bool'")
417 _set_data("has_obfuscation", domain, status)
418 logger.debug("EXIT!")