]> git.ipfire.org Git - ipfire.org.git/blobdiff - src/backend/wiki.py
wiki: Find other embedded files other than images
[ipfire.org.git] / src / backend / wiki.py
index 02bb24ff786ddbb3376b20fc57bd8d4d5578ffb5..9bd19eca9debd767a6a823332bf2278da358d6f8 100644 (file)
@@ -1,7 +1,11 @@
 #!/usr/bin/python3
 
 import difflib
+import hashlib
 import logging
+import markdown
+import markdown.extensions
+import markdown.preprocessors
 import os.path
 import re
 import urllib.parse
@@ -23,6 +27,14 @@ class Wiki(misc.Object):
                if res:
                        return Page(self.backend, res.id, data=res)
 
+       def __iter__(self):
+               return self._get_pages(
+                       "SELECT wiki.* FROM wiki_current current \
+                               LEFT JOIN wiki ON current.id = wiki.id \
+                               WHERE current.deleted IS FALSE \
+                               ORDER BY page",
+               )
+
        def make_path(self, page, path):
                # Nothing to do for absolute links
                if path.startswith("/"):
@@ -47,27 +59,23 @@ class Wiki(misc.Object):
                return page and not page.was_deleted()
 
        def get_page_title(self, page, default=None):
-               # Try to retrieve title from cache
-               title = self.memcache.get("wiki:title:%s" % page)
-               if title:
-                       return title
-
-               # If the title has not been in the cache, we will
-               # have to look it up
                doc = self.get_page(page)
                if doc:
                        title = doc.title
                else:
                        title = os.path.basename(page)
 
-               # Save in cache for forever
-               self.memcache.set("wiki:title:%s" % page, title)
-
                return title
 
        def get_page(self, page, revision=None):
                page = Page.sanitise_page_name(page)
-               assert page
+
+               # Split the path into parts
+               parts = page.split("/")
+
+               # Check if this is an action
+               if any((part.startswith("_") for part in parts)):
+                       return
 
                if revision:
                        return self._get_page("SELECT * FROM wiki WHERE page = %s \
@@ -97,8 +105,8 @@ class Wiki(misc.Object):
                page = self._get_page("INSERT INTO wiki(page, author_uid, markdown, changes, address) \
                        VALUES(%s, %s, %s, %s, %s) RETURNING *", page, author.uid, content or None, changes, address)
 
-               # Update cache
-               self.memcache.set("wiki:title:%s" % page.page, page.title)
+               # Store any linked files
+               page._store_linked_files()
 
                # Send email to all watchers
                page._send_watcher_emails(excludes=[author])
@@ -124,12 +132,10 @@ class Wiki(misc.Object):
                return ret
 
        def search(self, query, account=None, limit=None):
-               query = util.parse_search_query(query)
-
                res = self._get_pages("SELECT wiki.* FROM wiki_search_index search_index \
                        LEFT JOIN wiki ON search_index.wiki_id = wiki.id \
-                       WHERE search_index.document @@ to_tsquery('english', %s) \
-                               ORDER BY ts_rank(search_index.document, to_tsquery('english', %s)) DESC",
+                       WHERE search_index.document @@ websearch_to_tsquery('english', %s) \
+                               ORDER BY ts_rank(search_index.document, websearch_to_tsquery('english', %s)) DESC",
                        query, query)
 
                pages = []
@@ -154,13 +160,25 @@ class Wiki(misc.Object):
                self.db.execute("REFRESH MATERIALIZED VIEW wiki_search_index")
 
        def get_watchlist(self, account):
-               pages = self._get_pages(
-                       "WITH pages AS (SELECT * FROM wiki_current \
-                                       LEFT JOIN wiki ON wiki_current.id = wiki.id) \
-                       SELECT * FROM wiki_watchlist watchlist \
-                               LEFT JOIN pages ON watchlist.page = pages.page \
-                               WHERE watchlist.uid = %s",
-                       account.uid,
+               pages = self._get_pages("""
+                       WITH pages AS (
+                               SELECT
+                                       *
+                               FROM
+                                       wiki_current
+                               LEFT JOIN
+                                       wiki ON wiki_current.id = wiki.id
+                       )
+
+                       SELECT
+                               *
+                       FROM
+                               wiki_watchlist watchlist
+                       JOIN
+                               pages ON watchlist.page = pages.page
+                       WHERE
+                               watchlist.uid = %s
+                       """, account.uid,
                )
 
                return sorted(pages)
@@ -243,9 +261,7 @@ class Wiki(misc.Object):
                        filename, author.uid, address, mimetype, blob.id, len(data))
 
        def render(self, path, text):
-               r = WikiRenderer(self.backend, path)
-
-               return r.render(text)
+               return WikiRenderer(self.backend, path, text)
 
 
 class Page(misc.Object):
@@ -260,6 +276,8 @@ class Page(misc.Object):
                if isinstance(other, self.__class__):
                        return self.id == other.id
 
+               return NotImplemented
+
        def __lt__(self, other):
                if isinstance(other, self.__class__):
                        if self.page == other.page:
@@ -267,6 +285,8 @@ class Page(misc.Object):
 
                        return self.page < other.page
 
+               return NotImplemented
+
        @staticmethod
        def sanitise_page_name(page):
                if not page:
@@ -287,11 +307,11 @@ class Page(misc.Object):
 
        @property
        def url(self):
-               return self.page
+               return "/docs%s" % self.page
 
        @property
        def full_url(self):
-               return "https://wiki.ipfire.org%s" % self.url
+               return "https://www.ipfire.org%s" % self.url
 
        @property
        def page(self):
@@ -324,7 +344,30 @@ class Page(misc.Object):
 
        @property
        def html(self):
-               return self.backend.wiki.render(self.page, self.markdown)
+               lines = []
+
+               # Strip off the first line if it contains a heading (as it will be shown separately)
+               for i, line in enumerate(self.markdown.splitlines()):
+                       if i == 0 and line.startswith("#"):
+                               continue
+
+                       lines.append(line)
+
+               renderer = self.backend.wiki.render(self.page, "\n".join(lines))
+
+               return renderer.html
+
+       # Linked Files
+
+       @property
+       def files(self):
+               renderer = self.backend.wiki.render(self.page, self.markdown)
+
+               return renderer.files
+
+       def _store_linked_files(self):
+               self.db.executemany("INSERT INTO wiki_linked_files(page_id, path) \
+                       VALUES(%s, %s)", ((self.id, file) for file in self.files))
 
        @property
        def timestamp(self):
@@ -366,19 +409,6 @@ class Page(misc.Object):
        def check_acl(self, account):
                return self.backend.wiki.check_acl(self.page, account)
 
-       # Sidebar
-
-       @lazy_property
-       def sidebar(self):
-               parts = self.page.split("/")
-
-               while parts:
-                       sidebar = self.backend.wiki.get_page("%s/sidebar" % os.path.join(*parts))
-                       if sidebar:
-                               return sidebar
-
-                       parts.pop()
-
        # Watchers
 
        @lazy_property
@@ -447,9 +477,13 @@ class Page(misc.Object):
                        self.backend.messages.send_template("wiki/messages/page-changed",
                                account=watcher, page=self, priority=-10)
 
-       def restore(self, author, address):
+       def restore(self, author, address, comment=None):
                changes = "Restore to revision from %s" % self.timestamp.isoformat()
 
+               # Append comment
+               if comment:
+                       changes = "%s: %s" % (changes, comment)
+
                return self.backend.wiki.create_page(self.page,
                        author, self.markdown, changes=changes, address=address)
 
@@ -463,9 +497,11 @@ class File(misc.Object):
                if isinstance(other, self.__class__):
                        return self.id == other.id
 
+               return NotImplemented
+
        @property
        def url(self):
-               return os.path.join(self.path, self.filename)
+               return "/docs%s" % os.path.join(self.path, self.filename)
 
        @property
        def path(self):
@@ -493,9 +529,20 @@ class File(misc.Object):
                return self.data.created_at
 
        def delete(self, author=None):
+               if not self.can_be_deleted():
+                       raise RuntimeError("Cannot delete %s" % self)
+
                self.db.execute("UPDATE wiki_files SET deleted_at = NOW(), deleted_by = %s \
                        WHERE id = %s", author.uid if author else None, self.id)
 
+       def can_be_deleted(self):
+               # Cannot be deleted if still in use
+               if self.pages:
+                       return False
+
+               # Can be deleted
+               return True
+
        @property
        def deleted_at(self):
                return self.data.deleted_at
@@ -509,7 +556,7 @@ class File(misc.Object):
 
        def get_revisions(self):
                revisions = self.backend.wiki._get_files("SELECT * FROM wiki_files \
-                       WHERE path = %s ORDER BY created_at DESC", self.path)
+                       WHERE path = %s AND filename = %s ORDER BY created_at DESC", self.path, self.filename)
 
                return list(revisions)
 
@@ -519,6 +566,12 @@ class File(misc.Object):
        def is_image(self):
                return self.mimetype.startswith("image/")
 
+       def is_vector_image(self):
+               return self.mimetype in ("image/svg+xml",)
+
+       def is_bitmap_image(self):
+               return self.is_image() and not self.is_vector_image()
+
        @lazy_property
        def blob(self):
                res = self.db.get("SELECT data FROM wiki_blobs \
@@ -527,11 +580,18 @@ class File(misc.Object):
                if res:
                        return bytes(res.data)
 
-       def get_thumbnail(self, size):
-               cache_key = "-".join((self.path, util.normalize(self.filename), self.created_at.isoformat(), "%spx" % size))
+       async def get_thumbnail(self, size):
+               assert self.is_bitmap_image()
+
+               cache_key = "-".join((
+                       self.path,
+                       util.normalize(self.filename),
+                       self.created_at.isoformat(),
+                       "%spx" % size,
+               ))
 
                # Try to fetch the data from the cache
-               thumbnail = self.memcache.get(cache_key)
+               thumbnail = await self.backend.cache.get(cache_key)
                if thumbnail:
                        return thumbnail
 
@@ -539,10 +599,33 @@ class File(misc.Object):
                thumbnail = util.generate_thumbnail(self.blob, size)
 
                # Put it into the cache for forever
-               self.memcache.set(cache_key, thumbnail)
+               await self.backend.cache.set(cache_key, thumbnail)
 
                return thumbnail
 
+       @property
+       def pages(self):
+               """
+                       Returns a list of all pages this file is linked by
+               """
+               pages = self.backend.wiki._get_pages("""
+                       SELECT
+                               wiki.*
+                       FROM
+                               wiki_linked_files
+                       JOIN
+                               wiki_current ON wiki_linked_files.page_id = wiki_current.id
+                       LEFT JOIN
+                               wiki ON wiki_linked_files.page_id = wiki.id
+                       WHERE
+                               wiki_linked_files.path = %s
+                       ORDER BY
+                               wiki.page
+                       """, os.path.join(self.path, self.filename),
+               )
+
+               return list(pages)
+
 
 class WikiRenderer(misc.Object):
        schemas = (
@@ -557,17 +640,42 @@ class WikiRenderer(misc.Object):
        )
 
        # Links
-       links = re.compile(r"<a href=\"(.*?)\">(.*?)</a>")
+       _links = re.compile(r"<a href=\"(.*?)\">(.*?)</a>")
 
        # Images
-       images = re.compile(r"<img alt(?:=\"(.*?)\")? src=\"(.*?)\" (?:title=\"(.*?)\" )?/>")
+       _images = re.compile(r"<img alt(?:=\"(.*?)\")? src=\"(.*?)\" (?:title=\"(.*?)\" )?/>")
 
-       def init(self, path):
+       def init(self, path, text):
                self.path = path
+               self.text = text
+
+               # Markdown Renderer
+               self.renderer = markdown.Markdown(
+                       extensions=[
+                               LinkedFilesExtractorExtension(),
+                               PrettyLinksExtension(),
+                               "codehilite",
+                               "fenced_code",
+                               "footnotes",
+                               "nl2br",
+                               "sane_lists",
+                               "tables",
+                               "toc",
+                       ],
+               )
+
+               # Render!
+               self.html = self._render()
 
        def _render_link(self, m):
                url, text = m.groups()
 
+               # External Links
+               for schema in self.schemas:
+                       if url.startswith(schema):
+                               return """<a class="link-external" href="%s">%s</a>""" % \
+                                       (url, text or url)
+
                # Emails
                if "@" in url:
                        # Strip mailto:
@@ -577,26 +685,53 @@ class WikiRenderer(misc.Object):
                        return """<a class="link-external" href="mailto:%s">%s</a>""" % \
                                (url, text or url)
 
-               # External Links
-               for schema in self.schemas:
-                       if url.startswith(schema):
-                               return """<a class="link-external" href="%s">%s</a>""" % \
-                                       (url, text or url)
-
                # Everything else must be an internal link
                path = self.backend.wiki.make_path(self.path, url)
 
-               return """<a href="%s">%s</a>""" % \
+               return """<a href="/docs%s">%s</a>""" % \
                        (path, text or self.backend.wiki.get_page_title(path))
 
        def _render_image(self, m):
                alt_text, url, caption = m.groups()
 
+               # Compute a hash over the URL
+               h = hashlib.new("md5")
+               h.update(url.encode())
+               id = h.hexdigest()
+
+               html = """
+                       <div class="columns is-centered">
+                               <div class="column is-8">
+                                       <figure class="image modal-trigger" data-target="%(id)s">
+                                               <img src="/docs%(url)s" alt="%(caption)s">
+
+                                               <figcaption class="figure-caption">%(caption)s</figcaption>
+                                       </figure>
+
+                                       <div class="modal is-large" id="%(id)s">
+                                               <div class="modal-background"></div>
+
+                                               <div class="modal-content">
+                                                       <p class="image">
+                                                               <img src="/docs%(plain_url)s?s=1920" alt="%(caption)s"
+                                                                       loading="lazy">
+                                                       </p>
+                                               </div>
+
+                                               <button class="modal-close is-large" aria-label="close"></button>
+                                       </div>
+                               </div>
+                       </div>
+               """
+
                # Skip any absolute and external URLs
                if url.startswith("/") or url.startswith("https://") or url.startswith("http://"):
-                       return """<figure class="figure"><img src="%s" class="figure-img img-fluid rounded" alt="%s">
-                               <figcaption class="figure-caption">%s</figcaption></figure>
-                       """ % (url, alt_text, caption or "")
+                       return html % {
+                               "caption"   : caption or "",
+                               "id"        : id,
+                               "plain_url" : url,
+                               "url"       : url,
+                       }
 
                # Try to split query string
                url, delimiter, qs = url.partition("?")
@@ -605,7 +740,7 @@ class WikiRenderer(misc.Object):
                args = urllib.parse.parse_qs(qs)
 
                # Build absolute path
-               url = self.backend.wiki.make_path(self.path, url)
+               plain_url = url = self.backend.wiki.make_path(self.path, url)
 
                # Find image
                file = self.backend.wiki.get_file_by_path(url)
@@ -616,20 +751,99 @@ class WikiRenderer(misc.Object):
                if not "s" in args:
                        args["s"] = "920"
 
-               return """<figure class="figure"><img src="%s?%s" class="figure-img img-fluid rounded" alt="%s">
-               <figcaption class="figure-caption">%s</figcaption></figure>
-               """ % (url, urllib.parse.urlencode(args), caption, caption or "")
+               # Append arguments to the URL
+               if args:
+                       url = "%s?%s" % (url, urllib.parse.urlencode(args))
+
+               return html % {
+                       "caption"   : caption or "",
+                       "id"        : id,
+                       "plain_url" : plain_url,
+                       "url"       : url,
+               }
 
-       def render(self, text):
+       def _render(self):
                logging.debug("Rendering %s" % self.path)
 
-               # Borrow this from the blog
-               text = self.backend.blog._render_text(text, lang="markdown")
+               # Render...
+               text = self.renderer.convert(self.text)
 
                # Postprocess links
-               text = self.links.sub(self._render_link, text)
+               text = self._links.sub(self._render_link, text)
 
                # Postprocess images to <figure>
-               text = self.images.sub(self._render_image, text)
+               text = self._images.sub(self._render_image, text)
 
                return text
+
+       @lazy_property
+       def files(self):
+               """
+                       A list of all linked files that have been part of the rendered markup
+               """
+               files = []
+
+               for url in self.renderer.files:
+                       # Skip external images
+                       if url.startswith("https://") or url.startswith("http://"):
+                               continue
+
+                       # Make the URL absolute
+                       url = self.backend.wiki.make_path(self.path, url)
+
+                       # Check if this is a file (it could also just be a page)
+                       file = self.backend.wiki.get_file_by_path(url)
+                       if file:
+                               files.append(url)
+
+               return files
+
+
+class PrettyLinksExtension(markdown.extensions.Extension):
+       def extendMarkdown(self, md):
+               # Create links to Bugzilla
+               md.preprocessors.register(BugzillaLinksPreprocessor(md), "bugzilla", 10)
+
+               # Create links to CVE
+               md.preprocessors.register(CVELinksPreprocessor(md), "cve", 10)
+
+
+class BugzillaLinksPreprocessor(markdown.preprocessors.Preprocessor):
+       regex = re.compile(r"(?:#(\d{5,}))", re.I)
+
+       def run(self, lines):
+               for line in lines:
+                       yield self.regex.sub(r"[#\1](https://bugzilla.ipfire.org/show_bug.cgi?id=\1)", line)
+
+
+class CVELinksPreprocessor(markdown.preprocessors.Preprocessor):
+       regex = re.compile(r"(?:CVE)[\s\-](\d{4}\-\d+)")
+
+       def run(self, lines):
+               for line in lines:
+                       yield self.regex.sub(r"[CVE-\1](https://cve.mitre.org/cgi-bin/cvename.cgi?name=\1)", line)
+
+
+class LinkedFilesExtractor(markdown.treeprocessors.Treeprocessor):
+       """
+               Finds all Linked Files
+       """
+       def run(self, root):
+               self.md.files = []
+
+               # Find all images and store the URLs
+               for image in root.findall(".//img"):
+                       src = image.get("src")
+
+                       self.md.files.append(src)
+
+               # Find all links
+               for link in root.findall(".//a"):
+                       href = link.get("href")
+
+                       self.md.files.append(href)
+
+
+class LinkedFilesExtractorExtension(markdown.extensions.Extension):
+    def extendMarkdown(self, md):
+        md.treeprocessors.register(LinkedFilesExtractor(md), "linked-files-extractor", 10)