Ticket #1183: irclogsplugin-0.11-searching.patch

irclogs/__init__.py

@@ -11 +11 @@
-from trac.web.main import IRequestHandler
-from trac.util.html import escape, html, Markup
-from trac.util.text import to_unicode
+from trac.util.datefmt import parse_date, format_date
+from trac.search import ISearchSource
-
+import pyndexter
-
-class IrclogsPlugin(Component):
-    implements(INavigationContributor, ITemplateProvider, IRequestHandler, \
-
+, ISearchSource
-    _url_re = re.compile(r'^/irclogs(/(?P<year>\d{4})(/(?P<month>\d{2})'
-                         r'(/(?P<day>\d{2}))?)?)?/?$')
-# TODO: make the line format somewhat configurable
@@ -44 +47 @@
-                       doc='If not empty an button with this value as caption '
-                           'is added to the navigation bar, pointing to the '
-                           'irc plugin')
+    indexer = Option('irclogs', 'indexer', doc='pyndexter indexer URI')
-
-    def _to_unicode(self, iterable):
-        for line in iterable:
@@ -56 +60 @@
-              .replace('\\%d', '(?P<day>\d{2})')
-        )
-
+    def _get_file_glob(self):
+        if 0:
+            return pyndexter.util.quote(self.file_format) \
+                .replace('%25Y', '????') \
+                .replace('%25m', '??')   \
+                .replace('%25d', '??')
+        else: # workaround: globs are not working properly
+            return "*.log" 
+
-    def _get_filename(self, year, month, day):
-        return os.path.join(self.path, self.file_format
-                .replace('%Y', str(year))
@@ -268 +281 @@
-    def get_htdocs_dirs(self):
-        from pkg_resources import resource_filename
-        return [('irclogs', resource_filename(__name__, 'htdocs'))]
+
+    # ISearchSource methods
+    def get_search_filters(self, req):
+        if not req.perm.has_permission('IRCLOGS_VIEW'):
+            return []
+        return [('irclogs', 'IRC Logs', True),]
+
+    def _find_match(self, text, terms):
+        # fixme: why can't pyndexter give us the position(s) where it
+        # found the terms?
+        for line in text.splitlines():
+            for term in terms:
+                if term.lower() in line.lower():
+                    return line
+
+    def get_search_results(self, req, terms, filters):
+        """Return a list of search results matching each search term in `terms`.
+        
+        The `filters` parameters is a list of the enabled filters, each item
+        being the name of the tuples returned by `get_search_events`.
+
+        The events returned by this function must be tuples of the form
+        `(href, title, date, author, excerpt).`
+        """
+        if not 'irclogs' in filters:
+            return
+
+        framework = pyndexter.Framework(
+            indexer=self.indexer, stemmer='porter://', mode=pyndexter.READONLY)
+        framework.add_source('file://%s?include=%s' % (
+                pyndexter.util.quote(self.path), self._get_file_glob()))
+
+        for hit in framework.search(' '.join(terms)):
+            file = hit.uri.path.lstrip(self.path)
+            fileinfo = self._get_file_re().search(file).groupdict()
+
+            url = "/irclogs/%(year)s/%(month)s/%(day)s" % fileinfo
+            fragment = ''
+                
+            line = self._find_match(hit.current.content, terms)
+            if line:
+                lineinfo = self._line_re.search(line).groupdict()
+                fragment = "#%(time)s" % lineinfo
+                timestamp = parse_date("%(date)sT%(time)s" % lineinfo)
+            else:
+                timestamp = parse_date("%(year)s-%(month)s-%(day)s" % fileinfo)
+
+            title = "irclogs for %s" % format_date(timestamp)
+
+            yield (req.href(url)+fragment, title, timestamp, 'irclog', unicode(hit.excerpt(terms)))
+