fetch_blocks.py

   1 import reqto
   2 import time
   3 import bs4
   4 import fba
   5 import itertools
   6 import re
   7
   8 fba.c.execute(
   9     "SELECT domain, software FROM instances WHERE software IN ('pleroma', 'mastodon', 'friendica', 'misskey', 'gotosocial')"
  10 )
  11
  12 for blocker, software in fba.c.fetchall():
  13     print("DEBUG: blocker,software:", blocker, software)
  14     blockdict = []
  15     blocker = fba.tidyup(blocker)
  16     if software == "pleroma":
  17         print("INFO: blocker:", blocker)
  18         try:
  19             # Blocks
  20             federation = reqto.get(
  21                 f"https://{blocker}/nodeinfo/2.1.json", headers=fba.headers, timeout=5
  22             ).json()["metadata"]["federation"]
  23             if "mrf_simple" in federation:
  24                 for block_level, blocks in (
  25                     {**federation["mrf_simple"],
  26                     **{"quarantined_instances": federation["quarantined_instances"]}}
  27                 ).items():
  28                     for blocked in blocks:
  29                         print("DEBUG: BEFORE blocked:", blocked)
  30                         blocked = fba.tidyup(blocked)
  31                         print("DEBUG: AFTER blocked:", blocked)
  32
  33                         if blocked == "":
  34                             print("WARNING: blocked is empty after fba.tidyup():", blocker, block_level)
  35                             continue
  36
  37                         if blocked.count("*") > 1:
  38                             # -ACK!-oma also started obscuring domains without hash
  39                             fba.c.execute(
  40                                 "SELECT domain FROM instances WHERE domain LIKE ? ORDER BY rowid LIMIT 1", (blocked.replace("*", "_"),)
  41                             )
  42                             searchres = fba.c.fetchone()
  43                             print("DEBUG: searchres[]:", type(searchres))
  44                             if searchres != None:
  45                                 blocked = searchres[0]
  46
  47                         print("DEBUG: Looking up instance by domain:", blocked)
  48                         fba.c.execute(
  49                             "SELECT domain FROM instances WHERE domain = ?", (blocked,)
  50                         )
  51
  52                         if fba.c.fetchone() == None:
  53                             print("DEBUG: Domain wasn't found, adding:", blocked)
  54                             fba.add_instance(blocked)
  55
  56                         timestamp = int(time.time())
  57                         fba.c.execute(
  58                             "SELECT * FROM blocks WHERE blocker = ? AND blocked = ? AND block_level = ?",
  59                             (
  60                                blocker,
  61                                blocked,
  62                                block_level
  63                            ),
  64                         )
  65
  66                         if fba.c.fetchone() == None:
  67                             fba.block_instance(blocker, blocked, "unknown", block_level, timestamp, timestamp)
  68
  69                             if block_level == "reject":
  70                                 blockdict.append(
  71                                     {
  72                                         "blocked": blocked,
  73                                         "reason": None
  74                                     })
  75                         else:
  76                             fba.update_last_seen(timestamp, blocker, blocked, block_level)
  77
  78             fba.conn.commit()
  79
  80             # Reasons
  81             if "mrf_simple_info" in federation:
  82                 print("DEBUG: Found mrf_simple_info:", blocker)
  83                 for block_level, info in (
  84                     {**federation["mrf_simple_info"],
  85                     **(federation["quarantined_instances_info"]
  86                     if "quarantined_instances_info" in federation
  87                     else {})}
  88                 ).items():
  89                     print("DEBUG: block_level, info.items():", block_level, len(info.items()))
  90                     for blocked, reason in info.items():
  91                         print("DEBUG: BEFORE blocked:", blocked)
  92                         blocked = fba.tidyup(blocked)
  93                         print("DEBUG: AFTER blocked:", blocked)
  94
  95                         if blocked == "":
  96                             print("WARNING: blocked is empty after fba.tidyup():", blocker, block_level)
  97                             continue
  98
  99                         if blocked.count("*") > 1:
 100                             # same domain guess as above, but for reasons field
 101                             fba.c.execute(
 102                                 "SELECT domain FROM instances WHERE domain LIKE ? ORDER BY rowid LIMIT 1", (blocked.replace("*", "_"),)
 103                             )
 104                             searchres = fba.c.fetchone()
 105
 106                             if searchres != None:
 107                                 blocked = searchres[0]
 108
 109                         print("DEBUG: Updating block reason:", blocker, blocked, reason["reason"])
 110                         fba.update_block_reason(reason["reason"], blocker, blocked, block_level)
 111
 112                         for entry in blockdict:
 113                             if entry["blocked"] == blocked:
 114                                 print("DEBUG: Updating entry reason:", blocked)
 115                                 entry["reason"] = reason["reason"]
 116
 117             fba.conn.commit()
 118         except Exception as e:
 119             print("error:", e, blocker, software)
 120     elif software == "mastodon":
 121         print("INFO: blocker:", blocker)
 122         try:
 123             # json endpoint for newer mastodongs
 124             try:
 125                 json = {
 126                     "reject": [],
 127                     "media_removal": [],
 128                     "followers_only": [],
 129                     "report_removal": []
 130                 }
 131
 132                 # handling CSRF, I've saw at least one server requiring it to access the endpoint
 133                 print("DEBUG: Fetching meta:", blocker)
 134                 meta = bs4.BeautifulSoup(
 135                     reqto.get(f"https://{blocker}/about", headers=fba.headers, timeout=5).text,
 136                     "html.parser",
 137                 )
 138                 try:
 139                     csrf = meta.find("meta", attrs={"name": "csrf-token"})["content"]
 140                     print("DEBUG: Adding CSRF token:", blocker, csrf)
 141                     reqheaders = {**fba.headers, **{"x-csrf-token": csrf}}
 142                 except:
 143                     print("DEBUG: No CSRF token found, using normal headers:", blocker)
 144                     reqheaders = fba.headers
 145
 146                 print("DEBUG: Quering API domain_blocks:", blocker)
 147                 blocks = reqto.get(
 148                     f"https://{blocker}/api/v1/instance/domain_blocks", headers=reqheaders, timeout=5
 149                 ).json()
 150
 151                 print("DEBUG: blocks():", len(blocks))
 152                 for block in blocks:
 153                     entry = {'domain': block['domain'], 'hash': block['digest'], 'reason': block['comment']}
 154
 155                     print("DEBUG: severity,domain,hash,comment:", block['severity'], block['domain'], block['digest'], block['comment'])
 156                     if block['severity'] == 'suspend':
 157                         json['reject'].append(entry)
 158                     elif block['severity'] == 'silence':
 159                         json['followers_only'].append(entry)
 160                     elif block['severity'] == 'reject_media':
 161                         json['media_removal'].append(entry)
 162                     elif block['severity'] == 'reject_reports':
 163                         json['report_removal'].append(entry)
 164                     else:
 165                         print("WARNING: Unknown severity:", block['severity'], block['domain'])
 166             except:
 167                 print("DEBUG: Failed, Trying mastodon-specific fetches:", blocker)
 168                 json = fba.get_mastodon_blocks(blocker)
 169
 170             print("DEBUG: json.items():", blocker, len(json.items()))
 171             for block_level, blocks in json.items():
 172                 print("DEBUG: blocker,block_level,blocks():", blocker, block_level, len(blocks))
 173                 for instance in blocks:
 174                     blocked, blocked_hash, reason = instance.values()
 175                     print("DEBUG: blocked,hash,reason:", blocked, blocked_hash, reason)
 176
 177                     blocked = fba.tidyup(blocked)
 178                     print("DEBUG: blocked:", blocked)
 179
 180                     if blocked.count("*") < 1:
 181                         # No obsfucation for this instance
 182                         fba.c.execute(
 183                             "SELECT hash FROM instances WHERE domain = ? LIMIT 1", (blocked,)
 184                         )
 185
 186                         if fba.c.fetchone() == None:
 187                             print("DEBUG: Hash wasn't found, adding:", blocked)
 188                             fba.add_instance(blocked)
 189                     else:
 190                         # Doing the hash search for instance names as well to tidy up DB
 191                         fba.c.execute(
 192                             "SELECT domain FROM instances WHERE hash = ? LIMIT 1", (blocked_hash,)
 193                         )
 194                         searchres = fba.c.fetchone()
 195
 196                         if searchres != None:
 197                             print("DEBUG: Updating domain: ", searchres[0])
 198                             blocked = searchres[0]
 199
 200                     timestamp = int(time.time())
 201                     fba.c.execute(
 202                         "SELECT * FROM blocks WHERE blocker = ? AND blocked = ? AND block_level = ?",
 203                         (
 204                             blocker,
 205                             blocked if blocked.count("*") <= 1 else blocked_hash,
 206                             block_level
 207                         ),
 208                     )
 209
 210                     if fba.c.fetchone() == None:
 211                         fba.block_instance(blocker, blocked if blocked.count("*") <= 1 else blocked_hash, reason, block_level, timestamp, timestamp)
 212
 213                         if block_level == "reject":
 214                             blockdict.append(
 215                                 {
 216                                     "blocked": blocked,
 217                                     "reason": reason
 218                                 })
 219                     else:
 220                         fba.update_last_seen(timestamp, blocker, blocked if blocked.count("*") <= 1 else blocked_hash, block_level)
 221
 222                     if reason != '':
 223                         print("DEBUG: Updating block reason:", blocker, blocked, reason)
 224                         fba.update_block_reason(reason, blocker, blocked if blocked.count("*") <= 1 else blocked_hash, block_level)
 225
 226             fba.conn.commit()
 227         except Exception as e:
 228             print("error:", e, blocker, software)
 229     elif software == "friendica" or software == "misskey":
 230         print("INFO: blocker:", blocker)
 231         try:
 232             if software == "friendica":
 233                 json = fba.get_friendica_blocks(blocker)
 234             elif software == "misskey":
 235                 json = fba.get_misskey_blocks(blocker)
 236             for block_level, blocks in json.items():
 237                 for instance in blocks:
 238                     blocked, reason = instance.values()
 239                     blocked = fba.tidyup(blocked)
 240
 241                     print("BEFORE-blocked:", blocked)
 242                     if blocked.count("*") > 0:
 243                         # Some friendica servers also obscure domains without hash
 244                         fba.c.execute(
 245                             "SELECT domain FROM instances WHERE domain LIKE ? ORDER BY rowid LIMIT 1", (blocked.replace("*", "_"),)
 246                         )
 247                         searchres = fba.c.fetchone()
 248                         if searchres != None:
 249                             blocked = searchres[0]
 250
 251                     if blocked.count("?") > 0:
 252                         # Some obscure them with question marks, not sure if that's dependent on version or not
 253                         fba.c.execute(
 254                             "SELECT domain FROM instances WHERE domain LIKE ? ORDER BY rowid LIMIT 1", (blocked.replace("?", "_"),)
 255                         )
 256                         searchres = fba.c.fetchone()
 257                         if searchres != None:
 258                             blocked = searchres[0]
 259
 260                     print("AFTER-blocked:", blocked)
 261                     fba.c.execute(
 262                         "SELECT domain FROM instances WHERE domain = ?", (blocked,)
 263                     )
 264
 265                     if fba.c.fetchone() == None:
 266                         print("DEBUG: Hash wasn't found, adding:", blocked)
 267                         fba.add_instance(blocked)
 268
 269                     timestamp = int(time.time())
 270                     fba.c.execute(
 271                         "SELECT * FROM blocks WHERE blocker = ? AND blocked = ?",
 272                         (blocker, blocked),
 273                     )
 274                     if fba.c.fetchone() == None:
 275                         fba.block_instance(blocker, blocked, reason, block_level, timestamp, timestamp)
 276
 277                         if block_level == "reject":
 278                             blockdict.append(
 279                                 {
 280                                     "blocked": blocked,
 281                                     "reason": reason
 282                                 })
 283                     else:
 284                         fba.update_last_seen(timestamp, blocker, blocked, block_level)
 285
 286                     if reason != '':
 287                         print("DEBUG: Updating block reason:", blocker, blocked, reason)
 288                         fba.update_block_reason(reason, blocker, blocked, block_level)
 289
 290             fba.conn.commit()
 291         except Exception as e:
 292             print("error:", e, blocker, software)
 293     elif software == "gotosocial":
 294         print("INFO: blocker:", blocker)
 295         try:
 296             # Blocks
 297             federation = reqto.get(
 298                 f"https://{blocker}/api/v1/instance/peers?filter=suspended", headers=fba.headers, timeout=5
 299             ).json()
 300
 301             if (federation == None):
 302                 print("WARNING: No valid response:", blocker);
 303             else:
 304                 for peer in federation:
 305                     print("DEBUG: peer(),[]:", len(peer), type(peer))
 306                     if (isinstance(peer, str) and peer == "error"):
 307                         print("WARNING: Cannot continue, maybe authentication required?", blocker)
 308                         break
 309
 310                     blocked = peer["domain"].lower()
 311                     print("DEBUG: blocked:", blocked)
 312
 313                     if blocked.count("*") > 0:
 314                         # GTS does not have hashes for obscured domains, so we have to guess it
 315                         fba.c.execute(
 316                             "SELECT domain FROM instances WHERE domain LIKE ? ORDER BY rowid LIMIT 1", (blocked.replace("*", "_"),)
 317                         )
 318                         searchres = fba.c.fetchone()
 319
 320                         if searchres != None:
 321                             blocked = searchres[0]
 322
 323                     fba.c.execute(
 324                         "SELECT domain FROM instances WHERE domain = ?", (blocked,)
 325                     )
 326
 327                     if fba.c.fetchone() == None:
 328                         print("DEBUG: Hash wasn't found, adding:", blocked)
 329                         fba.add_instance(blocked)
 330
 331                     fba.c.execute(
 332                         "SELECT * FROM blocks WHERE blocker = ? AND blocked = ? AND block_level = ?",
 333                         (
 334                             blocker,
 335                             blocked,
 336                             "reject"
 337                         ),
 338                     )
 339                     timestamp = int(time.time())
 340
 341                     if fba.c.fetchone() == None:
 342                         fba.block_instance(blocker, blocked, "", "reject", timestamp, timestamp)
 343
 344                         blockdict.append(
 345                             {
 346                                 "blocked": blocked,
 347                                 "reason": None
 348                             })
 349                     else:
 350                         fba.update_last_seen(timestamp, blocker, blocked, "reject")
 351
 352                     if "public_comment" in peer:
 353                         reason = peer["public_comment"]
 354                         print("DEBUG: Updating block reason:", blocker, blocked, reason)
 355                         fba.update_block_reason(reason, blocker, blocked, "reject")
 356
 357                         for entry in blockdict:
 358                             if entry["blocked"] == blocked:
 359                                 entry["reason"] = reason
 360
 361                 fba.conn.commit()
 362         except Exception as e:
 363             print("error:", e, blocker, software)
 364     else:
 365         print("WARNING: Unknown software:", software)
 366
 367     if fba.config["bot_enabled"] and len(blockdict) > 0:
 368         send_bot_post(blocker, blockdict)
 369
 370     blockdict = []
 371
 372 fba.conn.close()