src/backend/wiki.py

   1 #!/usr/bin/python3
   2
   3 import difflib
   4 import logging
   5 import os.path
   6 import re
   7 import tornado.gen
   8 import urllib.parse
   9
  10 from . import misc
  11 from . import util
  12 from .decorators import *
  13
  14 class Wiki(misc.Object):
  15         def _get_pages(self, query, *args):
  16                 res = self.db.query(query, *args)
  17
  18                 for row in res:
  19                         yield Page(self.backend, row.id, data=row)
  20
  21         def _get_page(self, query, *args):
  22                 res = self.db.get(query, *args)
  23
  24                 if res:
  25                         return Page(self.backend, res.id, data=res)
  26
  27         def make_path(self, page, path):
  28                 # Nothing to do for absolute links
  29                 if path.startswith("/"):
  30                         pass
  31
  32                 # Relative links (one-level down)
  33                 elif path.startswith("./"):
  34                         path = os.path.join(page, path)
  35
  36                 # All other relative links
  37                 else:
  38                         p = os.path.dirname(page)
  39                         path = os.path.join(p, path)
  40
  41                 # Normalise links
  42                 return os.path.normpath(path)
  43
  44         def page_exists(self, path):
  45                 page = self.get_page(path)
  46
  47                 # Page must have been found and not deleted
  48                 return page and not page.was_deleted()
  49
  50         def get_page_title(self, page, default=None):
  51                 # Try to retrieve title from cache
  52                 title = self.memcache.get("wiki:title:%s" % page)
  53                 if title:
  54                         return title
  55
  56                 # If the title has not been in the cache, we will
  57                 # have to look it up
  58                 doc = self.get_page(page)
  59                 if doc:
  60                         title = doc.title
  61                 else:
  62                         title = os.path.basename(page)
  63
  64                 # Save in cache for forever
  65                 self.memcache.set("wiki:title:%s" % page, title)
  66
  67                 return title
  68
  69         def get_page(self, page, revision=None):
  70                 page = Page.sanitise_page_name(page)
  71                 assert page
  72
  73                 if revision:
  74                         return self._get_page("SELECT * FROM wiki WHERE page = %s \
  75                                 AND timestamp = %s", page, revision)
  76                 else:
  77                         return self._get_page("SELECT * FROM wiki WHERE page = %s \
  78                                 ORDER BY timestamp DESC LIMIT 1", page)
  79
  80         def get_recent_changes(self, account, limit=None):
  81                 pages = self._get_pages("SELECT * FROM wiki \
  82                         ORDER BY timestamp DESC")
  83
  84                 for page in pages:
  85                         if not page.check_acl(account):
  86                                 continue
  87
  88                         yield page
  89
  90                         limit -= 1
  91                         if not limit:
  92                                 break
  93
  94         def create_page(self, page, author, content, changes=None, address=None):
  95                 page = Page.sanitise_page_name(page)
  96
  97                 # Write page to the database
  98                 page = self._get_page("INSERT INTO wiki(page, author_uid, markdown, changes, address) \
  99                         VALUES(%s, %s, %s, %s, %s) RETURNING *", page, author.uid, content or None, changes, address)
 100
 101                 # Update cache
 102                 self.memcache.set("wiki:title:%s" % page.page, page.title)
 103
 104                 # Send email to all watchers
 105                 page._send_watcher_emails(excludes=[author])
 106
 107                 return page
 108
 109         def delete_page(self, page, author, **kwargs):
 110                 # Do nothing if the page does not exist
 111                 if not self.get_page(page):
 112                         return
 113
 114                 # Just creates a blank last version of the page
 115                 self.create_page(page, author=author, content=None, **kwargs)
 116
 117         def make_breadcrumbs(self, url):
 118                 # Split and strip all empty elements (double slashes)
 119                 parts = list(e for e in url.split("/") if e)
 120
 121                 ret = []
 122                 for part in ("/".join(parts[:i]) for i in range(1, len(parts))):
 123                         ret.append(("/%s" % part, self.get_page_title(part, os.path.basename(part))))
 124
 125                 return ret
 126
 127         def search(self, query, account=None, limit=None):
 128                 query = util.parse_search_query(query)
 129
 130                 res = self._get_pages("SELECT wiki.* FROM wiki_search_index search_index \
 131                         LEFT JOIN wiki ON search_index.wiki_id = wiki.id \
 132                         WHERE search_index.document @@ to_tsquery('english', %s) \
 133                                 ORDER BY ts_rank(search_index.document, to_tsquery('english', %s)) DESC",
 134                         query, query)
 135
 136                 pages = []
 137                 for page in res:
 138                         # Skip any pages the user doesn't have permission for
 139                         if not page.check_acl(account):
 140                                 continue
 141
 142                         # Return any other pages
 143                         pages.append(page)
 144
 145                         # Break when we have found enough pages
 146                         if limit and len(pages) >= limit:
 147                                 break
 148
 149                 return pages
 150
 151         def refresh(self):
 152                 """
 153                         Needs to be called after a page has been changed
 154                 """
 155                 self.db.execute("REFRESH MATERIALIZED VIEW wiki_search_index")
 156
 157         def get_watchlist(self, account):
 158                 pages = self._get_pages(
 159                         "WITH pages AS (SELECT * FROM wiki_current \
 160                                         LEFT JOIN wiki ON wiki_current.id = wiki.id) \
 161                         SELECT * FROM wiki_watchlist watchlist \
 162                                 LEFT JOIN pages ON watchlist.page = pages.page \
 163                                 WHERE watchlist.uid = %s",
 164                         account.uid,
 165                 )
 166
 167                 return sorted(pages)
 168
 169         # ACL
 170
 171         def check_acl(self, page, account):
 172                 res = self.db.query("SELECT * FROM wiki_acls \
 173                         WHERE %s ILIKE (path || '%%') ORDER BY LENGTH(path) DESC LIMIT 1", page)
 174
 175                 for row in res:
 176                         # Access not permitted when user is not logged in
 177                         if not account:
 178                                 return False
 179
 180                         # If user is in a matching group, we grant permission
 181                         for group in row.groups:
 182                                 if group in account.groups:
 183                                         return True
 184
 185                         # Otherwise access is not permitted
 186                         return False
 187
 188                 # If no ACLs are found, we permit access
 189                 return True
 190
 191         # Files
 192
 193         def _get_files(self, query, *args):
 194                 res = self.db.query(query, *args)
 195
 196                 for row in res:
 197                         yield File(self.backend, row.id, data=row)
 198
 199         def _get_file(self, query, *args):
 200                 res = self.db.get(query, *args)
 201
 202                 if res:
 203                         return File(self.backend, res.id, data=res)
 204
 205         def get_files(self, path):
 206                 files = self._get_files("SELECT * FROM wiki_files \
 207                         WHERE path = %s AND deleted_at IS NULL ORDER BY filename", path)
 208
 209                 return list(files)
 210
 211         def get_file_by_path(self, path):
 212                 path, filename = os.path.dirname(path), os.path.basename(path)
 213
 214                 return self._get_file("SELECT * FROM wiki_files \
 215                         WHERE path = %s AND filename = %s AND deleted_at IS NULL", path, filename)
 216
 217         def upload(self, path, filename, data, mimetype, author, address):
 218                 # Upload the blob first
 219                 blob = self.db.get("INSERT INTO wiki_blobs(data) VALUES(%s) RETURNING id", data)
 220
 221                 # Create entry for file
 222                 return self._get_file("INSERT INTO wiki_files(path, filename, author_uid, address, \
 223                         mimetype, blob_id, size) VALUES(%s,  %s, %s, %s, %s, %s, %s) RETURNING *", path,
 224                         filename, author.uid, address, mimetype, blob.id, len(data))
 225
 226         def render(self, path, text):
 227                 r = WikiRenderer(self.backend, path)
 228
 229                 return r.render(text)
 230
 231
 232 class Page(misc.Object):
 233         def init(self, id, data=None):
 234                 self.id = id
 235                 self.data = data
 236
 237         def __repr__(self):
 238                 return "<%s %s %s>" % (self.__class__.__name__, self.page, self.timestamp)
 239
 240         def __eq__(self, other):
 241                 if isinstance(other, self.__class__):
 242                         return self.id == other.id
 243
 244         def __lt__(self, other):
 245                 if isinstance(other, self.__class__):
 246                         if self.page == other.page:
 247                                 return self.timestamp < other.timestamp
 248
 249                         return self.page < other.page
 250
 251         @staticmethod
 252         def sanitise_page_name(page):
 253                 if not page:
 254                         return "/"
 255
 256                 # Make sure that the page name does NOT end with a /
 257                 if page.endswith("/"):
 258                         page = page[:-1]
 259
 260                 # Make sure the page name starts with a /
 261                 if not page.startswith("/"):
 262                         page = "/%s" % page
 263
 264                 # Remove any double slashes
 265                 page = page.replace("//", "/")
 266
 267                 return page
 268
 269         @property
 270         def url(self):
 271                 return self.page
 272
 273         @property
 274         def full_url(self):
 275                 return "https://wiki.ipfire.org%s" % self.url
 276
 277         @property
 278         def page(self):
 279                 return self.data.page
 280
 281         @property
 282         def title(self):
 283                 return self._title or os.path.basename(self.page[1:])
 284
 285         @property
 286         def _title(self):
 287                 if not self.markdown:
 288                         return
 289
 290                 # Find first H1 headline in markdown
 291                 markdown = self.markdown.splitlines()
 292
 293                 m = re.match(r"^#\s*(.*)( #)?$", markdown[0])
 294                 if m:
 295                         return m.group(1)
 296
 297         @lazy_property
 298         def author(self):
 299                 if self.data.author_uid:
 300                         return self.backend.accounts.get_by_uid(self.data.author_uid)
 301
 302         @property
 303         def markdown(self):
 304                 return self.data.markdown or ""
 305
 306         @property
 307         def html(self):
 308                 return self.backend.wiki.render(self.page, self.markdown)
 309
 310         @property
 311         def timestamp(self):
 312                 return self.data.timestamp
 313
 314         def was_deleted(self):
 315                 return self.markdown is None
 316
 317         @lazy_property
 318         def breadcrumbs(self):
 319                 return self.backend.wiki.make_breadcrumbs(self.page)
 320
 321         def get_latest_revision(self):
 322                 revisions = self.get_revisions()
 323
 324                 # Return first object
 325                 for rev in revisions:
 326                         return rev
 327
 328         def get_revisions(self):
 329                 return self.backend.wiki._get_pages("SELECT * FROM wiki \
 330                         WHERE page = %s ORDER BY timestamp DESC", self.page)
 331
 332         @lazy_property
 333         def previous_revision(self):
 334                 return self.backend.wiki._get_page("SELECT * FROM wiki \
 335                         WHERE page = %s AND timestamp < %s ORDER BY timestamp DESC \
 336                         LIMIT 1", self.page, self.timestamp)
 337
 338         @property
 339         def changes(self):
 340                 return self.data.changes
 341
 342         # ACL
 343
 344         def check_acl(self, account):
 345                 return self.backend.wiki.check_acl(self.page, account)
 346
 347         # Sidebar
 348
 349         @lazy_property
 350         def sidebar(self):
 351                 parts = self.page.split("/")
 352
 353                 while parts:
 354                         sidebar = self.backend.wiki.get_page("%s/sidebar" % os.path.join(*parts))
 355                         if sidebar:
 356                                 return sidebar
 357
 358                         parts.pop()
 359
 360         # Watchers
 361
 362         @lazy_property
 363         def diff(self):
 364                 if self.previous_revision:
 365                         diff = difflib.unified_diff(
 366                                 self.previous_revision.markdown.splitlines(),
 367                                 self.markdown.splitlines(),
 368                         )
 369
 370                         return "\n".join(diff)
 371
 372         @property
 373         def watchers(self):
 374                 res = self.db.query("SELECT uid FROM wiki_watchlist \
 375                         WHERE page = %s", self.page)
 376
 377                 for row in res:
 378                         # Search for account by UID and skip if none was found
 379                         account = self.backend.accounts.get_by_uid(row.uid)
 380                         if not account:
 381                                 continue
 382
 383                         # Return the account
 384                         yield account
 385
 386         def is_watched_by(self, account):
 387                 res = self.db.get("SELECT 1 FROM wiki_watchlist \
 388                         WHERE page = %s AND uid = %s", self.page, account.uid)
 389
 390                 if res:
 391                         return True
 392
 393                 return False
 394
 395         def add_watcher(self, account):
 396                 if self.is_watched_by(account):
 397                         return
 398
 399                 self.db.execute("INSERT INTO wiki_watchlist(page, uid) \
 400                         VALUES(%s, %s)", self.page, account.uid)
 401
 402         def remove_watcher(self, account):
 403                 self.db.execute("DELETE FROM wiki_watchlist \
 404                         WHERE page = %s AND uid = %s", self.page, account.uid)
 405
 406         def _send_watcher_emails(self, excludes=[]):
 407                 # Nothing to do if there was no previous revision
 408                 if not self.previous_revision:
 409                         return
 410
 411                 for watcher in self.watchers:
 412                         # Skip everyone who is excluded
 413                         if watcher in excludes:
 414                                 logging.debug("Excluding %s" % watcher)
 415                                 continue
 416
 417                         # Check permissions
 418                         if not self.backend.wiki.check_acl(self.page, watcher):
 419                                 logging.debug("Watcher %s does not have permissions" % watcher)
 420                                 continue
 421
 422                         logging.debug("Sending watcher email to %s" % watcher)
 423
 424                         # Compose message
 425                         self.backend.messages.send_template("wiki/messages/page-changed",
 426                                 sender="IPFire Wiki <wiki@ipfire.org>", recipients=[watcher],
 427                                 page=self, priority=-10)
 428
 429
 430 class File(misc.Object):
 431         def init(self, id, data):
 432                 self.id   = id
 433                 self.data = data
 434
 435         @property
 436         def url(self):
 437                 return os.path.join(self.path, self.filename)
 438
 439         @property
 440         def path(self):
 441                 return self.data.path
 442
 443         @property
 444         def filename(self):
 445                 return self.data.filename
 446
 447         @property
 448         def mimetype(self):
 449                 return self.data.mimetype
 450
 451         @property
 452         def size(self):
 453                 return self.data.size
 454
 455         @lazy_property
 456         def author(self):
 457                 if self.data.author_uid:
 458                         return self.backend.accounts.get_by_uid(self.data.author_uid)
 459
 460         @property
 461         def created_at(self):
 462                 return self.data.created_at
 463
 464         def is_pdf(self):
 465                 return self.mimetype in ("application/pdf", "application/x-pdf")
 466
 467         def is_image(self):
 468                 return self.mimetype.startswith("image/")
 469
 470         @lazy_property
 471         def blob(self):
 472                 res = self.db.get("SELECT data FROM wiki_blobs \
 473                         WHERE id = %s", self.data.blob_id)
 474
 475                 if res:
 476                         return bytes(res.data)
 477
 478         def get_thumbnail(self, size):
 479                 cache_key = "-".join((self.path, util.normalize(self.filename), self.created_at.isoformat(), "%spx" % size))
 480
 481                 # Try to fetch the data from the cache
 482                 thumbnail = self.memcache.get(cache_key)
 483                 if thumbnail:
 484                         return thumbnail
 485
 486                 # Generate the thumbnail
 487                 thumbnail = util.generate_thumbnail(self.blob, size)
 488
 489                 # Put it into the cache for forever
 490                 self.memcache.set(cache_key, thumbnail)
 491
 492                 return thumbnail
 493
 494
 495 class WikiRenderer(misc.Object):
 496         schemas = (
 497                 "ftp://",
 498                 "git://",
 499                 "http://",
 500                 "https://",
 501                 "rsync://",
 502                 "sftp://",
 503                 "ssh://",
 504                 "webcal://",
 505         )
 506
 507         # Links
 508         links = re.compile(r"<a href=\"(.*?)\">(.*?)</a>")
 509
 510         # Images
 511         images = re.compile(r"<img alt(?:=\"(.*?)\")? src=\"(.*?)\" (?:title=\"(.*?)\" )?/>")
 512
 513         def init(self, path):
 514                 self.path = path
 515
 516         def _render_link(self, m):
 517                 url, text = m.groups()
 518
 519                 # Emails
 520                 if "@" in url:
 521                         # Strip mailto:
 522                         if url.startswith("mailto:"):
 523                                 url = url[7:]
 524
 525                         return """<a class="link-external" href="mailto:%s">%s</a>""" % \
 526                                 (url, text or url)
 527
 528                 # External Links
 529                 for schema in self.schemas:
 530                         if url.startswith(schema):
 531                                 return """<a class="link-external" href="%s">%s</a>""" % \
 532                                         (url, text or url)
 533
 534                 # Everything else must be an internal link
 535                 path = self.backend.wiki.make_path(self.path, url)
 536
 537                 return """<a href="%s">%s</a>""" % \
 538                         (path, text or self.backend.wiki.get_page_title(path))
 539
 540         def _render_image(self, m):
 541                 alt_text, url, caption = m.groups()
 542
 543                 # Skip any absolute and external URLs
 544                 if url.startswith("/") or url.startswith("https://") or url.startswith("http://"):
 545                         return """<figure class="figure"><img src="%s" class="figure-img img-fluid rounded" alt="%s">
 546                                 <figcaption class="figure-caption">%s</figcaption></figure>
 547                         """ % (url, alt_text, caption or "")
 548
 549                 # Try to split query string
 550                 url, delimiter, qs = url.partition("?")
 551
 552                 # Parse query arguments
 553                 args = urllib.parse.parse_qs(qs)
 554
 555                 # Build absolute path
 556                 url = self.backend.wiki.make_path(self.path, url)
 557
 558                 # Find image
 559                 file = self.backend.wiki.get_file_by_path(url)
 560                 if not file or not file.is_image():
 561                         return "<!-- Could not find image %s in %s -->" % (url, self.path)
 562
 563                 # Scale down the image if not already done
 564                 if not "s" in args:
 565                         args["s"] = "920"
 566
 567                 return """<figure class="figure"><img src="%s?%s" class="figure-img img-fluid rounded" alt="%s">
 568                 <figcaption class="figure-caption">%s</figcaption></figure>
 569                 """ % (url, urllib.parse.urlencode(args), caption, caption or "")
 570
 571         def render(self, text):
 572                 logging.debug("Rendering %s" % self.path)
 573
 574                 # Borrow this from the blog
 575                 text = self.backend.blog._render_text(text, lang="markdown")
 576
 577                 # Postprocess links
 578                 text = self.links.sub(self._render_link, text)
 579
 580                 # Postprocess images to <figure>
 581                 text = self.images.sub(self._render_image, text)
 582
 583                 return text