src/backend/wiki.py

   1 #!/usr/bin/python3
   2
   3 import PIL
   4 import io
   5 import logging
   6 import os.path
   7 import re
   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 get_page_title(self, page, default=None):
  28                 doc = self.get_page(page)
  29                 if doc:
  30                         return doc.title
  31
  32                 return default or os.path.basename(page)
  33
  34         def get_page(self, page, revision=None):
  35                 page = Page.sanitise_page_name(page)
  36                 assert page
  37
  38                 if revision:
  39                         return self._get_page("SELECT * FROM wiki WHERE page = %s \
  40                                 AND timestamp = %s", page, revision)
  41                 else:
  42                         return self._get_page("SELECT * FROM wiki WHERE page = %s \
  43                                 ORDER BY timestamp DESC LIMIT 1", page)
  44
  45         def get_recent_changes(self, limit=None):
  46                 return self._get_pages("SELECT * FROM wiki \
  47                         WHERE timestamp >= NOW() - INTERVAL '4 weeks' \
  48                         ORDER BY timestamp DESC LIMIT %s", limit)
  49
  50         def create_page(self, page, author, content, changes=None, address=None):
  51                 page = Page.sanitise_page_name(page)
  52
  53                 return self._get_page("INSERT INTO wiki(page, author_uid, markdown, changes, address) \
  54                         VALUES(%s, %s, %s, %s, %s) RETURNING *", page, author.uid, content or None, changes, address)
  55
  56         def delete_page(self, page, author, **kwargs):
  57                 # Do nothing if the page does not exist
  58                 if not self.get_page(page):
  59                         return
  60
  61                 # Just creates a blank last version of the page
  62                 self.create_page(page, author=author, content=None, **kwargs)
  63
  64         def make_breadcrumbs(self, url):
  65                 # Split and strip all empty elements (double slashes)
  66                 parts = list(e for e in url.split("/") if e)
  67
  68                 ret = []
  69                 for part in ("/".join(parts[:i]) for i in range(1, len(parts))):
  70                         ret.append(("/%s" % part, self.get_page_title(part, os.path.basename(part))))
  71
  72                 return ret
  73
  74         def search(self, query, limit=None):
  75                 query = util.parse_search_query(query)
  76
  77                 res = self._get_pages("SELECT wiki.* FROM wiki_search_index search_index \
  78                         LEFT JOIN wiki ON search_index.wiki_id = wiki.id \
  79                         WHERE search_index.document @@ to_tsquery('english', %s) \
  80                                 ORDER BY ts_rank(search_index.document, to_tsquery('english', %s)) DESC \
  81                         LIMIT %s", query, query, limit)
  82
  83                 return list(res)
  84
  85         def refresh(self):
  86                 """
  87                         Needs to be called after a page has been changed
  88                 """
  89                 self.db.execute("REFRESH MATERIALIZED VIEW wiki_search_index")
  90
  91         # Files
  92
  93         def _get_files(self, query, *args):
  94                 res = self.db.query(query, *args)
  95
  96                 for row in res:
  97                         yield File(self.backend, row.id, data=row)
  98
  99         def _get_file(self, query, *args):
 100                 res = self.db.get(query, *args)
 101
 102                 if res:
 103                         return File(self.backend, res.id, data=res)
 104
 105         def get_files(self, path):
 106                 files = self._get_files("SELECT * FROM wiki_files \
 107                         WHERE path = %s AND deleted_at IS NULL ORDER BY filename", path)
 108
 109                 return list(files)
 110
 111         def get_file_by_path(self, path):
 112                 path, filename = os.path.dirname(path), os.path.basename(path)
 113
 114                 return self._get_file("SELECT * FROM wiki_files \
 115                         WHERE path = %s AND filename = %s AND deleted_at IS NULL", path, filename)
 116
 117         def upload(self, path, filename, data, mimetype, author, address):
 118                 # Upload the blob first
 119                 blob = self.db.get("INSERT INTO wiki_blobs(data) VALUES(%s) RETURNING id", data)
 120
 121                 # Create entry for file
 122                 return self._get_file("INSERT INTO wiki_files(path, filename, author_uid, address, \
 123                         mimetype, blob_id, size) VALUES(%s,  %s, %s, %s, %s, %s, %s) RETURNING *", path,
 124                         filename, author.uid, address, mimetype, blob.id, len(data))
 125
 126         def find_image(self, path, filename):
 127                 for p in (path, os.path.dirname(path)):
 128                         file = self.get_file_by_path(os.path.join(p, filename))
 129
 130                         if file and file.is_image():
 131                                 return file
 132
 133
 134 class Page(misc.Object):
 135         def init(self, id, data=None):
 136                 self.id = id
 137                 self.data = data
 138
 139         def __lt__(self, other):
 140                 if isinstance(other, self.__class__):
 141                         if self.page == other.page:
 142                                 return self.timestamp < other.timestamp
 143
 144                         return self.page < other.page
 145
 146         @staticmethod
 147         def sanitise_page_name(page):
 148                 if not page:
 149                         return "/"
 150
 151                 # Make sure that the page name does NOT end with a /
 152                 if page.endswith("/"):
 153                         page = page[:-1]
 154
 155                 # Make sure the page name starts with a /
 156                 if not page.startswith("/"):
 157                         page = "/%s" % page
 158
 159                 # Remove any double slashes
 160                 page = page.replace("//", "/")
 161
 162                 return page
 163
 164         @property
 165         def url(self):
 166                 return self.page
 167
 168         @property
 169         def page(self):
 170                 return self.data.page
 171
 172         @property
 173         def title(self):
 174                 return self._title or self.page[1:]
 175
 176         @property
 177         def _title(self):
 178                 if not self.markdown:
 179                         return
 180
 181                 # Find first H1 headline in markdown
 182                 markdown = self.markdown.splitlines()
 183
 184                 m = re.match(r"^# (.*)( #)?$", markdown[0])
 185                 if m:
 186                         return m.group(1)
 187
 188         @lazy_property
 189         def author(self):
 190                 if self.data.author_uid:
 191                         return self.backend.accounts.get_by_uid(self.data.author_uid)
 192
 193         def _render(self, text):
 194                 logging.debug("Rendering %s" % self)
 195
 196                 # Link images
 197                 replacements = []
 198                 for match in re.finditer(r"!\[(.*)\]\((.*)\)", text):
 199                         alt_text, url = match.groups()
 200
 201                         # Skip any absolute and external URLs
 202                         if url.startswith("/") or url.startswith("https://") or url.startswith("http://"):
 203                                 continue
 204
 205                         # Try to split query string
 206                         url, delimiter, qs = url.partition("?")
 207
 208                         # Parse query arguments
 209                         args = urllib.parse.parse_qs(qs)
 210
 211                         # Find image
 212                         file = self.backend.wiki.find_image(self.page, url)
 213                         if not file:
 214                                 continue
 215
 216                         # Scale down the image if not already done
 217                         if not "s" in args:
 218                                 args["s"] = "768"
 219
 220                         # Format URL
 221                         url = "%s?%s" % (file.url, urllib.parse.urlencode(args))
 222
 223                         replacements.append((match.span(), file, alt_text, url))
 224
 225                 # Apply all replacements
 226                 for (start, end), file, alt_text, url in reversed(replacements):
 227                         text = text[:start] + "[![%s](%s)](%s?action=detail)" % (alt_text, url, file.url) + text[end:]
 228
 229                 # Add wiki links
 230                 patterns = (
 231                         (r"\[\[([\w\d\/]+)(?:\|([\w\d\s]+))\]\]", r"/\1", r"\2", None, None),
 232                         (r"\[\[([\w\d\/\-]+)\]\]", r"/\1", r"\1", self.backend.wiki.get_page_title, r"\1"),
 233                 )
 234
 235                 for pattern, link, title, repl, args in patterns:
 236                         replacements = []
 237
 238                         for match in re.finditer(pattern, text):
 239                                 l = match.expand(link)
 240                                 t = match.expand(title)
 241
 242                                 if callable(repl):
 243                                         t = repl(match.expand(args)) or t
 244
 245                                 replacements.append((match.span(), t or l, l))
 246
 247                         # Apply all replacements
 248                         for (start, end), t, l in reversed(replacements):
 249                                 text = text[:start] + "[%s](%s)" % (t, l) + text[end:]
 250
 251                 # Borrow this from the blog
 252                 return self.backend.blog._render_text(text, lang="markdown")
 253
 254         @property
 255         def markdown(self):
 256                 return self.data.markdown
 257
 258         @property
 259         def html(self):
 260                 return self.data.html or self._render(self.markdown)
 261
 262         @property
 263         def timestamp(self):
 264                 return self.data.timestamp
 265
 266         def was_deleted(self):
 267                 return self.markdown is None
 268
 269         @lazy_property
 270         def breadcrumbs(self):
 271                 return self.backend.wiki.make_breadcrumbs(self.page)
 272
 273         def get_latest_revision(self):
 274                 revisions = self.get_revisions()
 275
 276                 # Return first object
 277                 for rev in revisions:
 278                         return rev
 279
 280         def get_revisions(self):
 281                 return self.backend.wiki._get_pages("SELECT * FROM wiki \
 282                         WHERE page = %s ORDER BY timestamp DESC", self.page)
 283
 284         @property
 285         def changes(self):
 286                 return self.data.changes
 287
 288         # Sidebar
 289
 290         @lazy_property
 291         def sidebar(self):
 292                 parts = self.page.split("/")
 293
 294                 while parts:
 295                         sidebar = self.backend.wiki.get_page("%s/sidebar" % os.path.join(*parts))
 296                         if sidebar:
 297                                 return sidebar
 298
 299                         parts.pop()
 300
 301
 302 class File(misc.Object):
 303         def init(self, id, data):
 304                 self.id   = id
 305                 self.data = data
 306
 307         @property
 308         def url(self):
 309                 return os.path.join(self.path, self.filename)
 310
 311         @property
 312         def path(self):
 313                 return self.data.path
 314
 315         @property
 316         def filename(self):
 317                 return self.data.filename
 318
 319         @property
 320         def mimetype(self):
 321                 return self.data.mimetype
 322
 323         @property
 324         def size(self):
 325                 return self.data.size
 326
 327         @lazy_property
 328         def author(self):
 329                 if self.data.author_uid:
 330                         return self.backend.accounts.get_by_uid(self.data.author_uid)
 331
 332         @property
 333         def created_at(self):
 334                 return self.data.created_at
 335
 336         def is_pdf(self):
 337                 return self.mimetype in ("application/pdf", "application/x-pdf")
 338
 339         def is_image(self):
 340                 return self.mimetype.startswith("image/")
 341
 342         @lazy_property
 343         def blob(self):
 344                 res = self.db.get("SELECT data FROM wiki_blobs \
 345                         WHERE id = %s", self.data.blob_id)
 346
 347                 if res:
 348                         return bytes(res.data)
 349
 350         def get_thumbnail(self, size):
 351                 cache_key = "-".join((self.path, util.normalize(self.filename), self.created_at.isoformat(), "%spx" % size))
 352
 353                 # Try to fetch the data from the cache
 354                 thumbnail = self.memcache.get(cache_key)
 355                 if thumbnail:
 356                         return thumbnail
 357
 358                 # Generate the thumbnail
 359                 thumbnail = self._generate_thumbnail(size)
 360
 361                 # Put it into the cache for forever
 362                 self.memcache.set(cache_key, thumbnail)
 363
 364                 return thumbnail
 365
 366         def _generate_thumbnail(self, size):
 367                 image = PIL.Image.open(io.BytesIO(self.blob))
 368
 369                 # Resize the image to the desired resolution
 370                 image.thumbnail((size, size), PIL.Image.ANTIALIAS)
 371
 372                 with io.BytesIO() as f:
 373                         # If writing out the image does not work with optimization,
 374                         # we try to write it out without any optimization.
 375                         try:
 376                                 image.save(f, image.format, optimize=True, quality=98)
 377                         except:
 378                                 image.save(f, image.format, quality=98)
 379
 380                         return f.getvalue()