[wip] A command which displays images from the article

wikicurses.1

@@ -35,6 +35,10 @@ G or end
 .RS 4
 Scroll to bottom
 .RE
+.PP
+:images
+.RS 4
+Launch a program (default: feh) with images from the article
 .SS Pager
 .PP
 c

wikicurses.conf.5

@@ -49,6 +49,17 @@ Hide the References section at the bottom of the page and strip citations from t
 .RE
 .SS keymap
 This section configures the keyboard bindings of wikicurses, in the format "key=command". Command can be any ex command supported by wikicurses.
+.SS images
+.PP
+program
+.RS 4
+The name of the command to use to show images. It should be able to take space separated URLs as arguments. Defaults to feh if unspecified.
+.RE
+.PP
+arguments
+.RS 4
+The list of arguments given to the image display program. The URLs are appended to this. This is a space separated list. Example: `--flag -abc -d --option-with-arg arg`.
+.RE
 .SS Other Sections
 Other sections are treated as wiki entries. The url is the url for api.php on the wiki. The username and password are required for editing.
 .PP

wikicurses/main.py

@@ -1,9 +1,11 @@
 import os
 import re
+import json
 import argparse
 import tempfile
 import subprocess
 import urllib.parse
+from functools import lru_cache
 
 import urwid
 
@@ -35,6 +37,111 @@ def tabComplete(text, matches):
     return match
 
 
+def showImages():
+    """
+    Launch program to display the images for the current article.
+
+    The program is specified in the config, along with it's args.
+    Otherwise, the default is `feh`.
+    """
+    if 'wikipedia.org' not in wiki.siteurl:
+        ex.notify('Cancelled: Only works on Wikipedia')
+        return
+
+    targets = fetchImageTargets(page.title)
+
+    if not targets:
+        ex.notify('No relevant images found')
+        return
+
+    ## We make a copy of the returned list here,
+    ## because it is memoized with `lru_cache` and we will be deleting
+    ## elements from it non-permanently.
+    image_info = list(fetchImageInfo(page.title))
+    filtered = []
+
+    for target in targets:
+        for index, i in enumerate(image_info):
+            title, url = i
+            if target in title:
+                filtered.append(url)
+                del image_info[index] # prevent more than one occurrence of image
+                break
+
+    try:
+        command = [settings.conf.get('images', 'program')]
+    except (settings.configparser.NoOptionError,
+            settings.configparser.NoSectionError):
+        command = ['feh']
+
+    try:
+        args = settings.conf.get('images', 'arguments').split(' ')
+    except (settings.configparser.NoOptionError,
+            settings.configparser.NoSectionError):
+        if command == ['feh']:
+            args = ['-q', '--scale-down', '--image-bg', 'white', '-g', '400x400']
+        else:
+            args = []
+
+    command.extend(args)
+    command.extend(filtered)
+
+    try:
+        with open(os.devnull, 'w') as fp:
+            subprocess.Popen(command, stdout=fp, stderr=fp)
+    except FileNotFoundError:
+        ex.notify('Program not found')
+        return
+
+
+@lru_cache()
+def fetchImageTargets(page_title):
+    """
+    Get filenames of relevant images from the current article.
+
+    These are used to filter out non-relevant images from the API result.
+    """
+    url = re.sub(r'api.php', 'index.php', wiki.siteurl)
+    page_title = re.sub(r' ', '_', page_title)
+    raw = wiki._query(customurl=url, action="raw", title=page_title)
+
+    ### Finding targets...
+    targets = []
+
+    ## of the form `[[Image:foobar.png]]`
+    for match in re.finditer(r'\b(?:File:|Image:)([^]|\n\r]+)', raw):
+        targets.append(match.group(1))
+
+    ## of the form `image1 = foobar.png`, `image2=foobar.png`,
+    ## and `image       =   bar baz.png` etc...
+    for match in re.finditer(
+            r'(image\d?[\t ]*?=[\t ]*)(.+?\..+?)((?=[^A-Za-z0-9\.])|(?:$))',
+            raw):
+        targets.append(match.group(2))
+
+    return targets
+
+
+@lru_cache()
+def fetchImageInfo(page_title):
+    """
+    Use API to fetch all image titles and urls on current article.
+
+    Returns a list of `(title, url)` tuples.
+    """
+    page_title = re.sub(r' ', '_', page_title)
+    result = wiki._query(action="query", titles=page_title, generator="images",
+                         prop="imageinfo", iiprop="url", format="json")
+
+    json_result = json.loads(result)
+    info = []
+
+    for v in json_result['query']['pages'].values():
+        info.append((v['title'], v['imageinfo'][0]['url']))
+
+    return info
+
+
 class SearchBox(urwid.Edit):
     title = "Search"
 
@@ -495,7 +602,7 @@ def cancel(button):
               'extlinks': Extlinks,
               'langs': Langs}
 cmds = tuple(overlaymap) + ('quit', 'bmark', 'open', 'edit', 'clearcache',
-                            'help', 'back', 'forward', 'random')
+                            'help', 'back', 'forward', 'random', 'images')
 
 def processCmd(cmd, *args):
     global current
@@ -528,6 +635,8 @@ def processCmd(cmd, *args):
             openPage(history[current], browsinghistory=True)
     elif cmd == 'random':
         openPage(wiki.random())
+    elif cmd == 'images':
+        showImages()
     elif cmd:
         ex.notify(cmd + ': Unknown Command')
 

wikicurses/wiki.py

@@ -87,7 +87,12 @@ def mainpage(self):
     def _query(self, post=False, **kwargs):
         params = {k: v for k, v in kwargs.items() if v is not False}
         data = urllib.parse.urlencode(params)
-        url = self.siteurl
+
+        try:
+            url = params['customurl']
+        except KeyError:
+            url = self.siteurl
+
         if post:
             data = data.encode()
         if not post: