boxes: Support autocomplete for stream + topic in message compose box.

This commit adds a preliminary support for autocomplete:
  `#**stream name>Topic name`

The current approach uses regex for identifying the last occurence of
required format. The dedicated autocomplete function is:
  `autocomplete_stream_and_topic`

A follow-up to this might be to handle '#**stream name**>' cases too.

Ticks a checkbox in #448.

zulipterminal/ui_tools/boxes.py

@@ -57,6 +57,10 @@
     from zulipterminal.model import Model
 
 
+# Example text: #**stream name>Topic name
+REGEX_AUTOCOMPLETE_STREAM_TOPIC = r"#\*\*([^*>]+)>([^*]*)$"
+
+
 class WriteBox(urwid.Pile):
     def __init__(self, view: Any) -> None:
         super().__init__(self.main_view(True))
@@ -368,6 +372,15 @@ def generic_autocomplete(self, text: str, state: Optional[int]
             prefix: text.rfind(prefix)
             for prefix in autocomplete_map
         }
+
+        # Matches autocomplete for stream + topic at the end of a text.
+        m = re.search(REGEX_AUTOCOMPLETE_STREAM_TOPIC, text)
+        if m:
+            autocomplete_map.update({
+                m.group(0): self.autocomplete_stream_and_topic
+            })
+            prefix_indices[m.group(0)] = m.start()
+
         found_prefix_indices = {
             prefix: index
             for prefix, index in prefix_indices.items()
@@ -484,6 +497,27 @@ def autocomplete_streams(self, text: str, prefix_string: str
                                     self.view.pinned_streams)
         return matched_data
 
+    def autocomplete_stream_and_topic(self, text: str, prefix_string: str
+                                      ) -> Tuple[List[str], List[str]]:
+        m = re.search(REGEX_AUTOCOMPLETE_STREAM_TOPIC, text)
+
+        stream = m.group(1) if m else ''
+
+        if self.model.is_valid_stream(stream):
+            stream_id = self.model.stream_id_from_name(stream)
+            topic_names = self.model.topics_in_stream(stream_id)
+        else:
+            topic_names = []
+
+        stream_prefix = f"#**{stream}>"
+        topic_suggestions = match_topics(topic_names,
+                                         text[len(stream_prefix):])
+
+        topic_typeaheads = format_string(topic_suggestions,
+                                         stream_prefix + '{}**')
+
+        return topic_typeaheads, topic_suggestions
+
     def autocomplete_emojis(self, text: str, prefix_string: str
                             ) -> Tuple[List[str], List[str]]:
         emoji_list = list(self.model.active_emoji_data.keys())