Added YouTube integration (yeah)

com.jeffser.Alpaca.json

@@ -85,6 +85,20 @@
 		}
 	    ]
 	},
+	{
+	    "name": "python3-pytube",
+	    "buildsystem": "simple",
+	    "build-commands": [
+		"pip3 install --verbose --exists-action=i --no-index --find-links=\"file://${PWD}\" --prefix=${FLATPAK_DEST} \"pytube\" --no-build-isolation"
+	    ],
+	    "sources": [
+		{
+		    "type": "file",
+		    "url": "https://files.pythonhosted.org/packages/51/64/bcf8632ed2b7a36bbf84a0544885ffa1d0b4bcf25cc0903dba66ec5fdad9/pytube-15.0.0-py3-none-any.whl",
+		    "sha256": "07b9904749e213485780d7eb606e5e5b8e4341aa4dccf699160876da00e12d78"
+		}
+	    ]
+	},
         {
             "name": "ollama",
             "buildsystem": "simple",

src/alpaca.gresource.xml

@@ -23,6 +23,7 @@
     <file alias="icons/scalable/status/brain-augemnted-symbolic.svg">icons/brain-augemnted-symbolic.svg</file>
     <file alias="icons/scalable/status/chain-link-loose-symbolic.svg">icons/chain-link-loose-symbolic.svg</file>
     <file alias="icons/scalable/status/document-text-symbolic.svg">icons/document-text-symbolic.svg</file>
+    <file alias="icons/scalable/status/play-symbolic.svg">icons/play-symbolic.svg</file>
     <file preprocess="xml-stripblanks">window.ui</file>
     <file preprocess="xml-stripblanks">gtk/help-overlay.ui</file>
   </gresource>

src/dialogs.py

@@ -182,8 +182,8 @@ def remove_attached_file_response(self, dialog, task, button):
 
 def remove_attached_file(self, button):
     dialog = Adw.AlertDialog(
-        heading=_("Remove File"),
+        heading=_("Remove Attachment"),
-        body=_("Are you sure you want to remove file?"),
+        body=_("Are you sure you want to remove attachment?"),
         close_response="cancel"
     )
     dialog.add_response("cancel", _("Cancel"))
@@ -290,3 +290,40 @@ def attach_file_response(self, file_dialog, result):
 def attach_file(self, filter):
     file_dialog = Gtk.FileDialog(default_filter=filter)
     file_dialog.open(self, None, lambda file_dialog, result: attach_file_response(self, file_dialog, result))
+
+
+# YouTube caption |
+
+def youtube_caption_response(self, dialog, task, video_url, caption_drop_down):
+    if dialog.choose_finish(task) == "accept":
+        buffer = self.message_text_view.get_buffer()
+        text = buffer.get_text(buffer.get_start_iter(), buffer.get_end_iter(), False).replace(video_url, "")
+        buffer.delete(buffer.get_start_iter(), buffer.get_end_iter())
+        buffer.insert(buffer.get_start_iter(), text, len(text))
+        self.attach_file("{}&caption_lang={}".format(video_url, caption_drop_down.get_selected_item().get_string().split(' | ')[1]), 'youtube')
+
+def youtube_caption(self, video_title, video_url, captions):
+    if len(captions) == 0:
+        self.show_toast("error", 9, self.main_overlay)
+        return
+    caption_list = Gtk.StringList()
+    for caption in captions: caption_list.append("{} | {}".format(caption.name, caption.code))
+    caption_drop_down = Gtk.DropDown(
+        enable_search=True,
+        model=caption_list
+    )
+    dialog = Adw.AlertDialog(
+        heading=_("Attach YouTube Video"),
+        body=_("{}\n\nPlease select a transcript to include").format(video_title),
+        extra_child=caption_drop_down,
+        close_response="cancel"
+    )
+    dialog.add_response("cancel", _("Cancel"))
+    dialog.add_response("accept", _("Accept"))
+    dialog.set_response_appearance("accept", Adw.ResponseAppearance.SUGGESTED)
+    dialog.choose(
+        parent = self,
+        cancellable = None,
+        callback = lambda dialog, task, video_url = video_url, caption_drop_down = caption_drop_down: youtube_caption_response(self, dialog, task, video_url, caption_drop_down)
+    )
+

src/icons/play-symbolic.svg

@@ -0,0 +1,2 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<svg xmlns="http://www.w3.org/2000/svg" height="16px" viewBox="0 0 16 16" width="16px"><path d="m 4.003906 4.070312 v 7.859376 c 0 1.070312 0.90625 1.066406 0.90625 1.066406 h 0.09375 c 0.171875 0 0.347656 -0.039063 0.5 -0.125 l 7 -4 c 0.308594 -0.171875 0.46875 -0.523438 0.46875 -0.875 c 0 -0.351563 -0.160156 -0.703125 -0.46875 -0.875 l -7 -4 c -0.152344 -0.085938 -0.328125 -0.125 -0.5 -0.125 h -0.09375 s -0.90625 0 -0.90625 1.074218 z m 0 0" fill="#222222"/></svg>

src/window.py

@@ -22,6 +22,7 @@ gi.require_version('GtkSource', '5')
 gi.require_version('GdkPixbuf', '2.0')
 from gi.repository import Adw, Gtk, Gdk, GLib, GtkSource, Gio, GdkPixbuf
 import json, requests, threading, os, re, base64, sys, gettext, locale, webbrowser, subprocess, uuid, shutil, tarfile, tempfile #, docx
+from pytube import YouTube
 from time import sleep
 from io import BytesIO
 from PIL import Image
@@ -122,7 +123,9 @@ class AlpacaWindow(Adw.ApplicationWindow):
             _("Cannot open image"),
             _("Cannot delete chat because it's the only one left"),
             _("There was an error with the local Ollama instance, so it has been reset"),
-            _("Image recognition is only available on specific models")
+            _("Image recognition is only available on specific models"),
+            _("This video does not have any transcriptions"),
+            _("This video is not available")
         ],
         "info": [
             _("Please select a model before chatting"),
@@ -182,10 +185,15 @@ class AlpacaWindow(Adw.ApplicationWindow):
         can_use_images = self.verify_if_image_can_be_used()
         for name, content in self.attachments.items():
             if content["type"] == 'image' and can_use_images: attached_images.append(name)
-            else: attached_files[name] = content['type']
+            else:
+                if content["type"] == 'youtube':
+                    attached_files[content['path']] = content['type']
+                else:
+                    attached_files[name] = content['type']
             if not os.path.exists(os.path.join(self.data_dir, "chats", self.chats['selected_chat'], id)):
                 os.makedirs(os.path.join(self.data_dir, "chats", self.chats['selected_chat'], id))
-            shutil.copy(content['path'], os.path.join(self.data_dir, "chats", self.chats['selected_chat'], id, name))
+            if content["type"] != 'youtube':
+                shutil.copy(content['path'], os.path.join(self.data_dir, "chats", self.chats['selected_chat'], id, name))
             content["button"].get_parent().remove(content["button"])
         self.attachments = {}
         self.attachment_box.set_visible(False)
@@ -420,7 +428,10 @@ class AlpacaWindow(Adw.ApplicationWindow):
             buffer = self.file_preview_text_view.get_buffer()
             buffer.delete(buffer.get_start_iter(), buffer.get_end_iter())
             buffer.insert(buffer.get_start_iter(), content, len(content))
-            self.file_preview_dialog.set_title(os.path.basename(file_path))
+            if file_type == 'youtube':
+                self.file_preview_dialog.set_title(YouTube(file_path).title)
+            else:
+                self.file_preview_dialog.set_title(os.path.basename(file_path))
             self.file_preview_dialog.present(self)
 
     def convert_history_to_ollama(self):
@@ -431,7 +442,10 @@ class AlpacaWindow(Adw.ApplicationWindow):
                 del new_message['files']
                 new_message['content'] = ''
                 for name, file_type in message['files'].items():
-                    file_path = os.path.join(self.data_dir, "chats", self.chats['selected_chat'], id, name)
+                    if file_type == 'youtube':
+                        file_path = name
+                    else:
+                        file_path = os.path.join(self.data_dir, "chats", self.chats['selected_chat'], id, name)
                     file_data = self.get_content_of_file(file_path, file_type)
                     if file_data: new_message['content'] += f"```[{name}]\n{file_data}\n```"
                 new_message['content'] += message['content']
@@ -531,11 +545,15 @@ class AlpacaWindow(Adw.ApplicationWindow):
                 child=file_container
             )
             for name, file_type in files.items():
-                shown_name='.'.join(name.split(".")[:-1])[:20] + (name[20:] and '..') + f".{name.split('.')[-1]}"
+                if file_type == 'youtube':
+                    yt = YouTube(name)
+                    shown_name=yt.title[:20] + (yt.title[20:] and '..')
+                else:
+                    shown_name='.'.join(name.split(".")[:-1])[:20] + (name[20:] and '..') + f".{name.split('.')[-1]}"
 
                 button_content = Adw.ButtonContent(
                     label=shown_name,
-                    icon_name="document-text-symbolic"
+                    icon_name="play-symbolic" if file_type=='youtube' else "document-text-symbolic"
                 )
                 button = Gtk.Button(
                     vexpand=False,
@@ -545,7 +563,11 @@ class AlpacaWindow(Adw.ApplicationWindow):
                     tooltip_text=name,
                     child=button_content
                 )
-                button.connect("clicked", lambda button, file_path=os.path.join(self.data_dir, "chats", self.chats['selected_chat'], id, name), file_type=file_type: self.preview_file(file_path, file_type))
+                if file_type == 'youtube':
+                    file_path = name
+                else:
+                    file_path = os.path.join(self.data_dir, "chats", self.chats['selected_chat'], id, name)
+                button.connect("clicked", lambda button, file_path=file_path, file_type=file_type: self.preview_file(file_path, file_type))
                 file_container.append(button)
             message_box.append(file_scroller)
 
@@ -1112,7 +1134,7 @@ class AlpacaWindow(Adw.ApplicationWindow):
         self.verify_connection()
 
     def get_content_of_file(self, file_path, file_type):
-        if not os.path.exists(file_path): return None
+        if file_type != 'youtube' and not os.path.exists(file_path): return None
         if file_type == 'image':
             try:
                 with Image.open(file_path) as img:
@@ -1141,6 +1163,12 @@ class AlpacaWindow(Adw.ApplicationWindow):
             for i, page in enumerate(reader.pages):
                 text += f"\n- Page {i}\n{page.extract_text()}\n"
             return text
+        elif file_type == 'youtube':
+            yt = YouTube(file_path)
+            text = "{}\n{}\n\n".format(yt.title, yt.author)
+            for event in yt.captions[file_path.split('&caption_lang=')[1]].json_captions['events']:
+                text += "{}\n".format(event['segs'][0]['utf8'].replace('\n', '\\n'))
+            return text
         #elif file_type == 'docx':
             #document = docx.Document(file_path)
             #if len(document.paragraphs) == 0: return None
@@ -1155,17 +1183,23 @@ class AlpacaWindow(Adw.ApplicationWindow):
         if len(self.attachments) == 0: self.attachment_box.set_visible(False)
 
     def attach_file(self, file_path, file_type):
-        name = self.generate_numbered_name(os.path.basename(file_path), self.attachments.keys())
+        if file_type == "youtube":
+            name = YouTube(file_path).title
+        else:
+            name = self.generate_numbered_name(os.path.basename(file_path), self.attachments.keys())
         content = self.get_content_of_file(file_path, file_type)
         if content:
-            shown_name='.'.join(name.split(".")[:-1])[:20] + (name[20:] and '..') + f".{name.split('.')[-1]}"
+            if file_type == "youtube":
-
+                shown_name=name[:20] + (name[20:] and '..')
+            else:
+                shown_name='.'.join(name.split(".")[:-1])[:20] + (name[20:] and '..') + f".{name.split('.')[-1]}"
             button_content = Adw.ButtonContent(
                 label=shown_name,
                 icon_name={
                     "image": "image-x-generic-symbolic",
                     "plain_text": "document-text-symbolic",
                     "pdf": "document-text-symbolic",
+                    "youtube": "play-symbolic",
                     #"docx": "document-text-symbolic"
                 }[file_type]
             )
@@ -1195,6 +1229,22 @@ class AlpacaWindow(Adw.ApplicationWindow):
         elif action_name == 'export_chat':
             self.export_chat(chat_name)
 
+    def text_received(self, clipboard, result):
+        text = clipboard.read_text_finish(result)
+        #Check if text is a Youtube URL
+        youtube_regex = re.compile(
+            r'(https?://)?(www\.)?(youtube|youtu|youtube-nocookie)\.(com|be)/'
+            r'(watch\?v=|embed/|v/|.+\?v=)?([^&=%\?]{11})')
+        if youtube_regex.match(text):
+            try:
+                yt = YouTube(text)
+                dialogs.youtube_caption(self, yt.title, text, yt.captions)
+            except Exception as e:
+                self.show_toast("error", 10, self.main_overlay)
+
+    def on_clipboard_paste(self, textview):
+        clipboard = Gdk.Display.get_default().get_clipboard()
+        clipboard.read_text_async(None, self.text_received)
 
     def __init__(self, **kwargs):
         super().__init__(**kwargs)
@@ -1214,6 +1264,7 @@ class AlpacaWindow(Adw.ApplicationWindow):
         self.get_application().create_action('delete_chat', self.chat_actions)
         self.get_application().create_action('rename_chat', self.chat_actions)
         self.get_application().create_action('export_chat', self.chat_actions)
+        self.message_text_view.connect("paste-clipboard", self.on_clipboard_paste)
         self.add_chat_button.connect("clicked", lambda button : self.new_chat())
         self.attachment_button.connect("clicked", lambda button, file_filter=self.file_filter_attachments: dialogs.attach_file(self, file_filter))
         self.create_model_name.get_delegate().connect("insert-text", self.check_alphanumeric)