Add caching for extracting text to search

src/pyj/read_book/db.pyj

@@ -360,7 +360,7 @@ class DB:
             mt = fdata.mimetype or 'application/octet-stream'
             if fdata.encoded:
                 result = Blob([base64decode(result)], {'type':mt})
-            proceed({'ok': True, 'result': result, 'name': name, 'mt': mt})
+            proceed({'ok': True, 'result': req.result, 'name': name, 'mt': mt})
         req.onerror = def(event):
             details = get_error_details(event)
             proceed({'ok': False, 'name': name, 'details': details})

src/pyj/read_book/search.pyj

@@ -53,14 +53,18 @@ class SearchOverlay:
             self._worker = start_worker('read_book.search')
         return self._worker
 
-    def queue_search(self, book, current_name):
+    def queue_search(self, query, book, current_name):
         spine = book.manifest.spine
         self.request_counter += 1
         self.worker.postMessage({
             'type': 'search', 'book_hash': book.book_hash, 'spine': spine, 'current_name': current_name,
-            'id': self.request_counter, 'stored_files': book.stored_files
+        'id': self.request_counter, 'stored_files': book.stored_files, 'query': query
         })
 
+    def clear_caches(self):
+        if self._worker:
+            self.worker.postMessage({'type': 'clear_caches'})
+
     def onkeydown(self, event):
         if event.key is 'Escape' or event.key is 'Esc':
             self.hide()

src/pyj/read_book/search_worker.pyj

@@ -3,6 +3,7 @@
 from __python__ import bound_methods, hash_literals
 
 from read_book.db import DB
+from read_book.resources import text_from_serialized_html
 
 GET_SPINE_FAILED = 1
 CONNECT_FAILED = 2
@@ -19,11 +20,15 @@ class Worker:
         self.searching = False
         self.current_query = None
         self.current_query_id = None
+        self.text_cache = {}
 
     @property
     def initialize_error_msg(self):
         return self.db?.initialize_error_msg
 
+    def clear_caches(self):
+        self.text_cache = {}
+
 
 wc = Worker()
 
@@ -33,11 +38,19 @@ def send_search_complete():
     wc.current_query = wc.current_query_id = None
 
 
+def search_in_text_of(name):
+    pass
+
+
 def queue_next_spine_item(spine_idx):
     name = wc.current_query.spine[spine_idx]
     if not name:
         send_search_complete()
         return
+    if wc.text_cache[name]:
+        search_in_text_of(name)
+        setTimeout(queue_next_spine_item.bind(None, spine_idx + 1), 0)
+        return
     query = wc.current_query
     wc.db.get_book_file(query.book_hash, query.stored_files, name, got_spine_item.bind(None, query.id, spine_idx))
 
@@ -46,7 +59,10 @@ def got_spine_item(query_id, spine_idx, result):
     if query_id is not wc.current_query_id:
         return
     if result.ok:
-        queue_next_spine_item(spine_idx + 1)
+        name = wc.current_query.spine[spine_idx]
+        wc.text_cache[name] = text_from_serialized_html(result.result)
+        search_in_text_of(name)
+        setTimeout(queue_next_spine_item.bind(None, spine_idx + 1), 0)
     else:
         if result.details is 'ENOENT':
             queue_next_spine_item(spine_idx + 1)
@@ -97,6 +113,8 @@ def worker_main():
                     perform_search(e.data)
             else:
                 wc.pending_search = e.data
+        elif e.data.type is 'clear_caches':
+            wc.clear_caches()
 
     self.onerror = def(e):
         send_error(UNHANDLED_ERROR, f'{e.line}:{e.message}')

src/pyj/read_book/view.pyj

@@ -932,6 +932,7 @@ class View:
             self.loaded_resources = {}
             self.content_popup_overlay.loaded_resources = {}
             self.timers.start_book(book)
+            self.search_overlay.clear_caches()
         unkey = username_key(get_interface_data().username)
         self.book = current_book.book = book
         hl = None