src/backend/wiki.py

   1 #!/usr/bin/python3
   2
   3 import PIL
   4 import difflib
   5 import io
   6 import logging
   7 import os.path
   8 import re
   9 import tornado.gen
  10 import urllib.parse
  11
  12 from . import misc
  13 from . import util
  14 from .decorators import *
  15
  16 class Wiki(misc.Object):
  17         def _get_pages(self, query, *args):
  18                 res = self.db.query(query, *args)
  19
  20                 for row in res:
  21                         yield Page(self.backend, row.id, data=row)
  22
  23         def _get_page(self, query, *args):
  24                 res = self.db.get(query, *args)
  25
  26                 if res:
  27                         return Page(self.backend, res.id, data=res)
  28
  29         def get_page_title(self, page, default=None):
  30                 doc = self.get_page(page)
  31                 if doc:
  32                         return doc.title
  33
  34                 return default or os.path.basename(page)
  35
  36         def get_page(self, page, revision=None):
  37                 page = Page.sanitise_page_name(page)
  38                 assert page
  39
  40                 if revision:
  41                         return self._get_page("SELECT * FROM wiki WHERE page = %s \
  42                                 AND timestamp = %s", page, revision)
  43                 else:
  44                         return self._get_page("SELECT * FROM wiki WHERE page = %s \
  45                                 ORDER BY timestamp DESC LIMIT 1", page)
  46
  47         def get_recent_changes(self, account, limit=None):
  48                 pages = self._get_pages("SELECT * FROM wiki \
  49                         WHERE timestamp >= NOW() - INTERVAL '4 weeks' \
  50                         ORDER BY timestamp DESC")
  51
  52                 for page in pages:
  53                         if not page.check_acl(account):
  54                                 continue
  55
  56                         yield page
  57
  58                         limit -= 1
  59                         if not limit:
  60                                 break
  61
  62         def create_page(self, page, author, content, changes=None, address=None):
  63                 page = Page.sanitise_page_name(page)
  64
  65                 # Write page to the database
  66                 page = self._get_page("INSERT INTO wiki(page, author_uid, markdown, changes, address) \
  67                         VALUES(%s, %s, %s, %s, %s) RETURNING *", page, author.uid, content or None, changes, address)
  68
  69                 # Send email to all watchers
  70                 page._send_watcher_emails(excludes=[author])
  71
  72                 return page
  73
  74         def delete_page(self, page, author, **kwargs):
  75                 # Do nothing if the page does not exist
  76                 if not self.get_page(page):
  77                         return
  78
  79                 # Just creates a blank last version of the page
  80                 self.create_page(page, author=author, content=None, **kwargs)
  81
  82         def make_breadcrumbs(self, url):
  83                 # Split and strip all empty elements (double slashes)
  84                 parts = list(e for e in url.split("/") if e)
  85
  86                 ret = []
  87                 for part in ("/".join(parts[:i]) for i in range(1, len(parts))):
  88                         ret.append(("/%s" % part, self.get_page_title(part, os.path.basename(part))))
  89
  90                 return ret
  91
  92         def search(self, query, account=None, limit=None):
  93                 query = util.parse_search_query(query)
  94
  95                 res = self._get_pages("SELECT wiki.* FROM wiki_search_index search_index \
  96                         LEFT JOIN wiki ON search_index.wiki_id = wiki.id \
  97                         WHERE search_index.document @@ to_tsquery('english', %s) \
  98                                 ORDER BY ts_rank(search_index.document, to_tsquery('english', %s)) DESC",
  99                         query, query)
 100
 101                 for page in res:
 102                         # Skip any pages the user doesn't have permission for
 103                         if not page.check_acl(account):
 104                                 continue
 105
 106                         # Return any other pages
 107                         yield page
 108
 109                         limit -= 1
 110                         if not limit:
 111                                 break
 112
 113         def refresh(self):
 114                 """
 115                         Needs to be called after a page has been changed
 116                 """
 117                 self.db.execute("REFRESH MATERIALIZED VIEW wiki_search_index")
 118
 119         # ACL
 120
 121         def check_acl(self, page, account):
 122                 res = self.db.query("SELECT * FROM wiki_acls \
 123                         WHERE %s ILIKE (path || '%%') ORDER BY LENGTH(path) DESC LIMIT 1", page)
 124
 125                 for row in res:
 126                         # Access not permitted when user is not logged in
 127                         if not account:
 128                                 return False
 129
 130                         # If user is in a matching group, we grant permission
 131                         for group in row.groups:
 132                                 if group in account.groups:
 133                                         return True
 134
 135                         # Otherwise access is not permitted
 136                         return False
 137
 138                 # If no ACLs are found, we permit access
 139                 return True
 140
 141         # Files
 142
 143         def _get_files(self, query, *args):
 144                 res = self.db.query(query, *args)
 145
 146                 for row in res:
 147                         yield File(self.backend, row.id, data=row)
 148
 149         def _get_file(self, query, *args):
 150                 res = self.db.get(query, *args)
 151
 152                 if res:
 153                         return File(self.backend, res.id, data=res)
 154
 155         def get_files(self, path):
 156                 files = self._get_files("SELECT * FROM wiki_files \
 157                         WHERE path = %s AND deleted_at IS NULL ORDER BY filename", path)
 158
 159                 return list(files)
 160
 161         def get_file_by_path(self, path):
 162                 path, filename = os.path.dirname(path), os.path.basename(path)
 163
 164                 return self._get_file("SELECT * FROM wiki_files \
 165                         WHERE path = %s AND filename = %s AND deleted_at IS NULL", path, filename)
 166
 167         def upload(self, path, filename, data, mimetype, author, address):
 168                 # Upload the blob first
 169                 blob = self.db.get("INSERT INTO wiki_blobs(data) VALUES(%s) RETURNING id", data)
 170
 171                 # Create entry for file
 172                 return self._get_file("INSERT INTO wiki_files(path, filename, author_uid, address, \
 173                         mimetype, blob_id, size) VALUES(%s,  %s, %s, %s, %s, %s, %s) RETURNING *", path,
 174                         filename, author.uid, address, mimetype, blob.id, len(data))
 175
 176         def find_image(self, path, filename):
 177                 for p in (path, os.path.dirname(path)):
 178                         file = self.get_file_by_path(os.path.join(p, filename))
 179
 180                         if file and file.is_image():
 181                                 return file
 182
 183
 184 class Page(misc.Object):
 185         def init(self, id, data=None):
 186                 self.id = id
 187                 self.data = data
 188
 189         def __repr__(self):
 190                 return "<%s %s %s>" % (self.__class__.__name__, self.page, self.timestamp)
 191
 192         def __eq__(self, other):
 193                 if isinstance(other, self.__class__):
 194                         return self.id == other.id
 195
 196         def __lt__(self, other):
 197                 if isinstance(other, self.__class__):
 198                         if self.page == other.page:
 199                                 return self.timestamp < other.timestamp
 200
 201                         return self.page < other.page
 202
 203         @staticmethod
 204         def sanitise_page_name(page):
 205                 if not page:
 206                         return "/"
 207
 208                 # Make sure that the page name does NOT end with a /
 209                 if page.endswith("/"):
 210                         page = page[:-1]
 211
 212                 # Make sure the page name starts with a /
 213                 if not page.startswith("/"):
 214                         page = "/%s" % page
 215
 216                 # Remove any double slashes
 217                 page = page.replace("//", "/")
 218
 219                 return page
 220
 221         @property
 222         def url(self):
 223                 return self.page
 224
 225         @property
 226         def full_url(self):
 227                 return "https://wiki.ipfire.org%s" % self.url
 228
 229         @property
 230         def page(self):
 231                 return self.data.page
 232
 233         @property
 234         def title(self):
 235                 return self._title or os.path.basename(self.page[1:])
 236
 237         @property
 238         def _title(self):
 239                 if not self.markdown:
 240                         return
 241
 242                 # Find first H1 headline in markdown
 243                 markdown = self.markdown.splitlines()
 244
 245                 m = re.match(r"^# (.*)( #)?$", markdown[0])
 246                 if m:
 247                         return m.group(1)
 248
 249         @lazy_property
 250         def author(self):
 251                 if self.data.author_uid:
 252                         return self.backend.accounts.get_by_uid(self.data.author_uid)
 253
 254         def _render(self, text):
 255                 logging.debug("Rendering %s" % self)
 256
 257                 # Link images
 258                 replacements = []
 259                 for match in re.finditer(r"!\[(.*?)\]\((.*?)\)", text):
 260                         alt_text, url = match.groups()
 261
 262                         # Skip any absolute and external URLs
 263                         if url.startswith("/") or url.startswith("https://") or url.startswith("http://"):
 264                                 continue
 265
 266                         # Try to split query string
 267                         url, delimiter, qs = url.partition("?")
 268
 269                         # Parse query arguments
 270                         args = urllib.parse.parse_qs(qs)
 271
 272                         # Find image
 273                         file = self.backend.wiki.find_image(self.page, url)
 274                         if not file:
 275                                 continue
 276
 277                         # Scale down the image if not already done
 278                         if not "s" in args:
 279                                 args["s"] = "768"
 280
 281                         # Format URL
 282                         url = "%s?%s" % (file.url, urllib.parse.urlencode(args))
 283
 284                         replacements.append((match.span(), file, alt_text, url))
 285
 286                 # Apply all replacements
 287                 for (start, end), file, alt_text, url in reversed(replacements):
 288                         text = text[:start] + "[![%s](%s)](%s?action=detail)" % (alt_text, url, file.url) + text[end:]
 289
 290                 # Add wiki links
 291                 patterns = (
 292                         (r"\[\[([\w\d\/\-\.]+)(?:\|(.+?))\]\]", r"\1", r"\2", None, True),
 293                         (r"\[\[([\w\d\/\-\.]+)\]\]", r"\1", r"\1", self.backend.wiki.get_page_title, True),
 294
 295                         # External links
 296                         (r"\[\[((?:ftp|git|https?|rsync|sftp|ssh|webcal)\:\/\/.+?)(?:\|(.+?))\]\]",
 297                                 r"\1", r"\2", None, False),
 298
 299                         # Mail
 300                         (r"\[\[([a-zA-Z0-9_.+-]+@[a-zA-Z0-9-]+\.[a-zA-Z0-9-.]+)\]\]",
 301                                 r"\1", r"\1", None, False),
 302                 )
 303
 304                 for pattern, link, title, repl, internal in patterns:
 305                         replacements = []
 306
 307                         for match in re.finditer(pattern, text):
 308                                 l = match.expand(link)
 309                                 t = match.expand(title)
 310
 311                                 if internal:
 312                                         # Allow relative links
 313                                         if not l.startswith("/"):
 314                                                 l = os.path.join(self.page, l)
 315
 316                                         # Normalise links
 317                                         l = os.path.normpath(l)
 318
 319                                 if callable(repl):
 320                                         t = repl(l) or t
 321
 322                                 replacements.append((match.span(), t or l, l))
 323
 324                         # Apply all replacements
 325                         for (start, end), t, l in reversed(replacements):
 326                                 text = text[:start] + "[%s](%s)" % (t, l) + text[end:]
 327
 328                 # Borrow this from the blog
 329                 return self.backend.blog._render_text(text, lang="markdown")
 330
 331         @property
 332         def markdown(self):
 333                 return self.data.markdown or ""
 334
 335         @property
 336         def html(self):
 337                 return self._render(self.markdown)
 338
 339         @property
 340         def timestamp(self):
 341                 return self.data.timestamp
 342
 343         def was_deleted(self):
 344                 return self.markdown is None
 345
 346         @lazy_property
 347         def breadcrumbs(self):
 348                 return self.backend.wiki.make_breadcrumbs(self.page)
 349
 350         def get_latest_revision(self):
 351                 revisions = self.get_revisions()
 352
 353                 # Return first object
 354                 for rev in revisions:
 355                         return rev
 356
 357         def get_revisions(self):
 358                 return self.backend.wiki._get_pages("SELECT * FROM wiki \
 359                         WHERE page = %s ORDER BY timestamp DESC", self.page)
 360
 361         @lazy_property
 362         def previous_revision(self):
 363                 return self.backend.wiki._get_page("SELECT * FROM wiki \
 364                         WHERE page = %s AND timestamp < %s ORDER BY timestamp DESC \
 365                         LIMIT 1", self.page, self.timestamp)
 366
 367         @property
 368         def changes(self):
 369                 return self.data.changes
 370
 371         # ACL
 372
 373         def check_acl(self, account):
 374                 return self.backend.wiki.check_acl(self.page, account)
 375
 376         # Sidebar
 377
 378         @lazy_property
 379         def sidebar(self):
 380                 parts = self.page.split("/")
 381
 382                 while parts:
 383                         sidebar = self.backend.wiki.get_page("%s/sidebar" % os.path.join(*parts))
 384                         if sidebar:
 385                                 return sidebar
 386
 387                         parts.pop()
 388
 389         # Watchers
 390
 391         @lazy_property
 392         def diff(self):
 393                 if self.previous_revision:
 394                         diff = difflib.unified_diff(
 395                                 self.previous_revision.markdown.splitlines(),
 396                                 self.markdown.splitlines(),
 397                         )
 398
 399                         return "\n".join(diff)
 400
 401         @property
 402         def watchers(self):
 403                 res = self.db.query("SELECT uid FROM wiki_watchlist \
 404                         WHERE page = %s", self.page)
 405
 406                 for row in res:
 407                         # Search for account by UID and skip if none was found
 408                         account = self.backend.accounts.get_by_uid(row.uid)
 409                         if not account:
 410                                 continue
 411
 412                         # Return the account
 413                         yield account
 414
 415         def is_watched_by(self, account):
 416                 res = self.db.get("SELECT 1 FROM wiki_watchlist \
 417                         WHERE page = %s AND uid = %s", self.page, account.uid)
 418
 419                 if res:
 420                         return True
 421
 422                 return False
 423
 424         def add_watcher(self, account):
 425                 if self.is_watched_by(account):
 426                         return
 427
 428                 self.db.execute("INSERT INTO wiki_watchlist(page, uid) \
 429                         VALUES(%s, %s)", self.page, account.uid)
 430
 431         def remove_watcher(self, account):
 432                 self.db.execute("DELETE FROM wiki_watchlist \
 433                         WHERE page = %s AND uid = %s", self.page, account.uid)
 434
 435         def _send_watcher_emails(self, excludes=[]):
 436                 # Nothing to do if there was no previous revision
 437                 if not self.previous_revision:
 438                         return
 439
 440                 for watcher in self.watchers:
 441                         # Skip everyone who is excluded
 442                         if watcher in excludes:
 443                                 logging.debug("Excluding %s" % watcher)
 444                                 continue
 445
 446                         logging.debug("Sending watcher email to %s" % watcher)
 447
 448                         # Compose message
 449                         self.backend.messages.send_template("wiki/messages/page-changed",
 450                                 recipients=[watcher], page=self, priority=-10)
 451
 452
 453 class File(misc.Object):
 454         def init(self, id, data):
 455                 self.id   = id
 456                 self.data = data
 457
 458         @property
 459         def url(self):
 460                 return os.path.join(self.path, self.filename)
 461
 462         @property
 463         def path(self):
 464                 return self.data.path
 465
 466         @property
 467         def filename(self):
 468                 return self.data.filename
 469
 470         @property
 471         def mimetype(self):
 472                 return self.data.mimetype
 473
 474         @property
 475         def size(self):
 476                 return self.data.size
 477
 478         @lazy_property
 479         def author(self):
 480                 if self.data.author_uid:
 481                         return self.backend.accounts.get_by_uid(self.data.author_uid)
 482
 483         @property
 484         def created_at(self):
 485                 return self.data.created_at
 486
 487         def is_pdf(self):
 488                 return self.mimetype in ("application/pdf", "application/x-pdf")
 489
 490         def is_image(self):
 491                 return self.mimetype.startswith("image/")
 492
 493         @lazy_property
 494         def blob(self):
 495                 res = self.db.get("SELECT data FROM wiki_blobs \
 496                         WHERE id = %s", self.data.blob_id)
 497
 498                 if res:
 499                         return bytes(res.data)
 500
 501         def get_thumbnail(self, size):
 502                 cache_key = "-".join((self.path, util.normalize(self.filename), self.created_at.isoformat(), "%spx" % size))
 503
 504                 # Try to fetch the data from the cache
 505                 thumbnail = self.memcache.get(cache_key)
 506                 if thumbnail:
 507                         return thumbnail
 508
 509                 # Generate the thumbnail
 510                 thumbnail = self._generate_thumbnail(size)
 511
 512                 # Put it into the cache for forever
 513                 self.memcache.set(cache_key, thumbnail)
 514
 515                 return thumbnail
 516
 517         def _generate_thumbnail(self, size):
 518                 image = PIL.Image.open(io.BytesIO(self.blob))
 519
 520                 # Resize the image to the desired resolution
 521                 image.thumbnail((size, size), PIL.Image.ANTIALIAS)
 522
 523                 with io.BytesIO() as f:
 524                         # If writing out the image does not work with optimization,
 525                         # we try to write it out without any optimization.
 526                         try:
 527                                 image.save(f, image.format, optimize=True, quality=98)
 528                         except:
 529                                 image.save(f, image.format, quality=98)
 530
 531                         return f.getvalue()