pygpt-net 2.4.41__py3-none-any.whl → 2.4.44__py3-none-any.whl

pygpt_net/core/attachments/context.py

@@ -6,7 +6,7 @@
 # GitHub:  https://github.com/szczyglis-dev/py-gpt   #
 # MIT License                                        #
 # Created By  : Marcin Szczygliński                  #
-# Updated Date: 2024.12.14 08:00:00                  #
+# Updated Date: 2024.12.16 01:00:00                  #
 # ================================================== #
 
 import copy
@@ -15,7 +15,9 @@ import shutil
 import uuid
 
 from shutil import copyfile
-from typing import Optional, List, Dict, Any
+from typing import Optional, List, Dict, Any, Tuple
+
+from llama_index.core import Document
 
 from pygpt_net.core.bridge import BridgeContext
 from pygpt_net.core.events import KernelEvent
@@ -125,7 +127,7 @@ class Context:
         meta_path = self.get_dir(meta)
         context = ""
         if os.path.exists(meta_path) and os.path.isdir(meta_path):
-            for file in meta.additional_ctx:
+            for file in meta.get_additional_ctx():
                 if ("type" not in file
                         or file["type"] not in ["local_file", "url"]):
                     continue
@@ -183,7 +185,7 @@ class Context:
 
         indexed = False
         # index files if not indexed by auto_index
-        for i, file in enumerate(meta.additional_ctx):
+        for i, file in enumerate(meta.get_additional_ctx()):
             if "indexed" not in file or not file["indexed"]:
                 file_id = file["uuid"]
                 file_idx_path = os.path.join(meta_path, file_id)
@@ -313,7 +315,7 @@ class Context:
             prompt: str,
             auto_index: bool = False,
             real_path: Optional[str] = None
-    ) -> dict:
+    ) -> Dict[str, Any]:
         """
         Upload attachment for context
 
@@ -325,7 +327,10 @@ class Context:
         :return: Dict with attachment data
         """
         if self.is_verbose():
-            print("Uploading for meta ID: {}".format(meta.id))
+            if meta.group:
+                print("Uploading for meta group ID: {}".format(meta.group.id))
+            else:
+                print("Uploading for meta ID: {}".format(meta.id))
 
         # prepare idx dir
         name = os.path.basename(attachment.path)
@@ -342,11 +347,18 @@ class Context:
             if auto_index:
                 print("Attachments: vector index path: {}".format(index_path))
 
-        # store content to read
-        src_file = self.store_content(attachment, file_idx_path)
+        documents = None
+        
+        # store content to read, and get docs if type == web
+        src_file, docs = self.store_content(attachment, file_idx_path)
+        if attachment.type == AttachmentItem.TYPE_URL:
+            documents = docs
+
+        # extract text content using data loader, and get docs if type == file
+        content, docs = self.read_content(attachment, src_file, prompt)
+        if attachment.type == AttachmentItem.TYPE_FILE:
+            documents = docs
 
-        # extract text content using data loader
-        content = self.read_content(attachment, src_file, prompt)
         if content:
             text_path = os.path.join(file_idx_path, file_id + ".txt")
             with open(text_path, "w", encoding="utf-8") as f:
@@ -372,7 +384,7 @@ class Context:
             source = src_file
             if attachment.type == AttachmentItem.TYPE_URL:
                 source = attachment.path  # URL
-            doc_ids = self.index_attachment(attachment.type, source, index_path)
+            doc_ids = self.index_attachment(attachment.type, source, index_path, documents=documents)
 
         result = {
             "name": name,
@@ -402,57 +414,61 @@ class Context:
             attachment: AttachmentItem,
             path: str,
             prompt: str
-    ) -> str:
+    ) -> Tuple[str, List[Document]]:
         """
         Read content from attachment
 
         :param attachment: AttachmentItem instance
         :param path: source file path
         :param prompt: user input prompt
-        :return: content
+        :return: text content, list of documents
         """
         content = ""
+        docs = []
         if attachment.type == AttachmentItem.TYPE_FILE:
             loader_kwargs = {
                 "prompt": prompt,
             }  # extra loader kwargs
-            content = self.window.core.idx.indexing.read_text_content(
+            content, docs = self.window.core.idx.indexing.read_text_content(
                 path=path,
                 loader_kwargs=loader_kwargs,
             )
         elif attachment.type == AttachmentItem.TYPE_URL:
-            # directly from path
+            # directly from before stored path
             with open(path, "r", encoding="utf-8") as f:
-                content = f.read()  # already crawled
-
-        return content
+                content = f.read()  # data is already crawled in `store_content`
+        return content, docs
 
     def store_content(
             self,
             attachment: AttachmentItem,
             dir: str
-    ) -> str:
+    ) -> Tuple[str, List[Document]]:
         """
         Prepare content for attachment
 
         :param attachment: AttachmentItem instance
         :param dir: directory to save content
-        :return: content
+        :return: path, list of documents
         """
         path = None
+        docs = []
         if attachment.type == AttachmentItem.TYPE_FILE:
-            # copy raw file
+            # copy raw file only
             name = os.path.basename(attachment.path)
             path = os.path.join(dir, name)
             if os.path.exists(path):
                 os.remove(path)
             copyfile(attachment.path, path)
         elif attachment.type == AttachmentItem.TYPE_URL:
-            web_type = self.window.core.idx.indexing.get_webtype(attachment.path)
-            content = self.window.core.idx.indexing.read_web_content(
-                url=attachment.path,
-                type=web_type,  # webpage, default, TODO: add more types
-                extra_args={},
+            loader = attachment.extra.get("loader")
+            input_params = attachment.extra.get("input_params")
+            input_config = attachment.extra.get("input_config")
+            self.window.core.idx.indexing.update_loader_args(loader, input_config)  # update config
+            content, docs = self.window.core.idx.indexing.read_web_content(
+                url="",
+                type=loader,
+                extra_args=input_params,
             )
             # src file save
             name = "url.txt"
@@ -461,14 +477,14 @@ class Context:
                 os.remove(path)
             with open(path, "w", encoding="utf-8") as f:
                 f.write(content)
-        return path
+        return path, docs
 
     def index_attachment(
             self,
             type: str,
             source: str,
             idx_path: str,
-            documents: Optional[list] = None
+            documents: Optional[List[Document]] = None
     ) -> list:
         """
         Index attachment
@@ -496,7 +512,7 @@ class Context:
         :param meta: CtxMeta instance
         :return: list of attachments
         """
-        return meta.additional_ctx
+        return meta.get_additional_ctx()
 
     def get_dir(self, meta: CtxMeta) -> str:
         """
@@ -506,6 +522,8 @@ class Context:
         :return: directory path
         """
         meta_uuid = str(meta.uuid)
+        if meta.group:
+            meta_uuid = str(meta.group.uuid)
         return os.path.join(self.window.core.config.get_user_dir("ctx_idx"), meta_uuid)
 
     def get_selected_model(self, mode: str = "summary"):
@@ -562,7 +580,7 @@ class Context:
         :param meta: CtxMeta instance
         :return: number of attachments
         """
-        return len(meta.additional_ctx)
+        return len(meta.get_additional_ctx())
 
     def delete(
             self,
@@ -577,11 +595,11 @@ class Context:
         :param item: Attachment item dict
         :param delete_files: delete files
         """
-        meta.additional_ctx.remove(item)
+        meta.remove_additional_ctx(item)
         self.window.core.ctx.save(meta.id)
         if delete_files:
             self.delete_local(meta, item)
-        if len(meta.additional_ctx) == 0:
+        if len(meta.get_additional_ctx()) == 0:
             self.delete_index(meta)
 
     def delete_by_meta(self, meta: CtxMeta):
@@ -613,7 +631,7 @@ class Context:
         :param meta: CtxMeta instance
         :param delete_files: delete files
         """
-        meta.additional_ctx = []
+        meta.reset_additional_ctx()
         self.window.core.ctx.save(meta.id)
         if delete_files:
             self.delete_index(meta)
@@ -644,7 +662,7 @@ class Context:
         :param meta: CtxMeta instance
         :param delete_files: delete files
         """
-        meta.additional_ctx = []
+        meta.reset_additional_ctx()
         self.window.core.ctx.save(meta.id)
         if delete_files:
             self.delete_index(meta)

pygpt_net/core/audio/__init__.py

@@ -6,7 +6,7 @@
 # GitHub:  https://github.com/szczyglis-dev/py-gpt   #
 # MIT License                                        #
 # Created By  : Marcin Szczygliński                  #
-# Updated Date: 2024.12.14 08:00:00                  #
+# Updated Date: 2024.12.14 18:00:00                  #
 # ================================================== #
 
 import re
@@ -17,6 +17,8 @@ from bs4 import UnicodeDammit
 from pygpt_net.provider.audio_input.base import BaseProvider as InputBaseProvider
 from pygpt_net.provider.audio_output.base import BaseProvider as OutputBaseProvider
 
+from .whisper import Whisper
+
 
 class Audio:
     def __init__(self, window=None):
@@ -26,6 +28,7 @@ class Audio:
         :param window: Window instance
         """
         self.window = window
+        self.whisper = Whisper(window)
         self.providers = {
             "input": {},
             "output": {},

pygpt_net/core/audio/whisper.py

@@ -0,0 +1,37 @@
+#!/usr/bin/env python3
+# -*- coding: utf-8 -*-
+# ================================================== #
+# This file is a part of PYGPT package               #
+# Website: https://pygpt.net                         #
+# GitHub:  https://github.com/szczyglis-dev/py-gpt   #
+# MIT License                                        #
+# Created By  : Marcin Szczygliński                  #
+# Updated Date: 2024.12.14 18:00:00                  #
+# ================================================== #
+
+from typing import List
+
+class Whisper:
+    def __init__(self, window=None):
+        """
+        Whisper core
+
+        :param window: Window instance
+        """
+        self.window = window
+        self.voices = [
+            "alloy",
+            "echo",
+            "fable",
+            "onyx",
+            "nova",
+            "shimmer",
+        ]
+
+    def get_voices(self) -> List[str]:
+        """
+        Get whisper voices
+
+        :return: whisper voice name
+        """
+        return self.voices

pygpt_net/core/bridge/worker.py

@@ -60,14 +60,14 @@ class BridgeWorker(QObject, QRunnable):
                     extra=self.extra,
                 )
 
-            # Llama-index: chat with files
+            # LlamaIndex: chat with files
             elif self.mode == MODE_LLAMA_INDEX:
                 result = self.window.core.idx.chat.call(
                     context=self.context,
                     extra=self.extra,
                 )
 
-            # Llama-index: agents
+            # LlamaIndex: agents
             elif self.mode == MODE_AGENT_LLAMA:
                 result = self.window.core.agents.runner.call(
                     context=self.context,

pygpt_net/core/db/__init__.py

@@ -122,6 +122,7 @@ class Database:
             'updated_ts',
             'created_ts',
             'uuid',
+            'additional_ctx_json',
         ]
         columns["idx_ctx"] = [
             'id',
@@ -242,7 +243,7 @@ class Database:
                 'sort_by': columns["ctx_group"],
                 'search_fields': ['id', 'name'],
                 'timestamp_columns': ['created_ts', 'updated_ts'],
-                'json_columns': [],
+                'json_columns': ['additional_ctx_json'],
                 'default_sort': 'id',
                 'default_order': 'DESC',
                 'primary_key': 'id',

pygpt_net/core/debug/attachments.py

@@ -38,6 +38,7 @@ class AttachmentsDebug:
                     'mode': mode,
                     'type': attachment.type,
                     'consumed': attachment.consumed,
+                    'extra': attachment.extra,
                 }
                 self.window.core.debug.add(self.id, attachment.name, str(data))
 

pygpt_net/core/debug/events.py

@@ -6,10 +6,17 @@
 # GitHub:  https://github.com/szczyglis-dev/py-gpt   #
 # MIT License                                        #
 # Created By  : Marcin Szczygliński                  #
-# Updated Date: 2024.11.20 03:00:00                  #
+# Updated Date: 2024.12.14 19:00:00                  #
 # ================================================== #
 
-from pygpt_net.core.events import ControlEvent, AppEvent, Event, KernelEvent, RenderEvent
+from pygpt_net.core.events import (
+    ControlEvent,
+    AppEvent,
+    Event,
+    KernelEvent,
+    RenderEvent,
+    BaseEvent
+)
 
 
 class EventsDebug:
@@ -25,19 +32,24 @@ class EventsDebug:
     def update(self):
         """Update debug window."""
         self.window.core.debug.begin(self.id)
-        self.window.core.debug.add(self.id, 'App Events:', str(AppEvent.__dict__))
-        self.window.core.debug.add(self.id, 'Control Events:', str(ControlEvent.__dict__))
-        self.window.core.debug.add(self.id, 'Kernel Events:', str(KernelEvent.__dict__))
-        self.window.core.debug.add(self.id, 'Render Events:', str(RenderEvent.__dict__))
-        self.window.core.debug.add(self.id, 'Plugin Events:', str(self.extract_plugins()))
+        self.window.core.debug.add(self.id, 'App Events:', str(self.extract_events(AppEvent)))
+        self.window.core.debug.add(self.id, 'Control Events:', str(self.extract_events(ControlEvent)))
+        self.window.core.debug.add(self.id, 'Kernel Events:', str(self.extract_events(KernelEvent)))
+        self.window.core.debug.add(self.id, 'Render Events:', str(self.extract_events(RenderEvent)))
+        self.window.core.debug.add(self.id, 'Plugin Events:', str(self.extract_events(Event)))
+        self.window.core.debug.add(self.id, '----', '')
         self.window.core.debug.add(self.id, 'Voice Cmds (all):', str(self.window.core.access.voice.commands))
         self.window.core.debug.add(self.id, 'Voice Cmds (allowed):', str(self.window.core.access.voice.get_commands()))
         self.window.core.debug.end(self.id)
 
-    def extract_plugins(self) -> dict:
-        """Extract plugin events"""
+    def extract_events(self, events: BaseEvent) -> dict:
+        """
+        Extract events
+
+        :param events: Events class
+        """
         result = {}
-        for property, value in vars(Event).items():
+        for property, value in vars(events).items():
             if (isinstance(value, str)
                     or isinstance(value, int)
                     or isinstance(value, float)

pygpt_net/core/debug/tabs.py

@@ -6,7 +6,7 @@
 # GitHub:  https://github.com/szczyglis-dev/py-gpt   #
 # MIT License                                        #
 # Created By  : Marcin Szczygliński                  #
-# Updated Date: 2024.12.12 01:00:00                  #
+# Updated Date: 2024.12.14 17:00:00                  #
 # ================================================== #
 
 class TabsDebug:
@@ -22,10 +22,12 @@ class TabsDebug:
     def update(self):
         """Update debug window."""
         self.window.core.debug.begin(self.id)
-        self.window.core.debug.add(self.id, 'current PID', str(self.window.controller.ui.tabs.get_current_pid()))
+        self.window.core.debug.add(self.id, 'current Col', str(self.window.controller.ui.tabs.get_current_column_idx()))
         self.window.core.debug.add(self.id, 'current IDX', str(self.window.controller.ui.tabs.get_current_idx()))
+        self.window.core.debug.add(self.id, 'current Tab', str(self.window.controller.ui.tabs.get_current_tab()))
+        self.window.core.debug.add(self.id, 'current PID', str(self.window.controller.ui.tabs.get_current_pid()))
         self.window.core.debug.add(self.id, 'current Type', str(self.window.controller.ui.tabs.get_current_type()))
-        self.window.core.debug.add(self.id, 'current Column', str(self.window.controller.ui.tabs.get_current_column_idx()))
+        self.window.core.debug.add(self.id, '----', '')
         self.window.core.debug.add(self.id, 'last_pid', str(self.window.core.tabs.last_pid))
         self.window.core.debug.add(self.id, 'locked', str(self.window.controller.ui.tabs.locked))
         self.window.core.debug.add(self.id, 'col', str(self.window.controller.ui.tabs.col))
@@ -40,6 +42,7 @@ class TabsDebug:
             self.window.core.debug.add(self.id, "PID ["+str(pid)+"]", str(data))
 
         # mapping PID => meta.id
+        self.window.core.debug.add(self.id, '----', '')
         self.window.core.debug.add(self.id, 'PID => meta.id', str(self.window.core.ctx.output.mapping))
         self.window.core.debug.add(self.id, '(last) meta.id => PID', str(self.window.core.ctx.output.last_pids))
         self.window.core.debug.add(self.id, '(last) PID', str(self.window.core.ctx.output.last_pid))

pygpt_net/core/history.py

@@ -6,11 +6,12 @@
 # GitHub:  https://github.com/szczyglis-dev/py-gpt   #
 # MIT License                                        #
 # Created By  : Marcin Szczygliński                  #
-# Updated Date: 2024.01.19 05:00:00                  #
+# Updated Date: 2024.12.14 22:00:00                  #
 # ================================================== #
 
 import datetime
 import os
+from typing import List
 
 from pygpt_net.item.ctx import CtxItem
 from pygpt_net.provider.core.history.txt_file import TxtFileProvider
@@ -46,7 +47,7 @@ class History:
         # delete all txt history files from history dir
         self.provider.truncate()
 
-    def remove_items(self, items: list):
+    def remove_items(self, items: List[CtxItem]):
         """
         Remove items from history (txt files)
 

pygpt_net/core/idx/__init__.py

@@ -6,7 +6,7 @@
 # GitHub:  https://github.com/szczyglis-dev/py-gpt   #
 # MIT License                                        #
 # Created By  : Marcin Szczygliński                  #
-# Updated Date: 2024.12.14 08:00:00                  #
+# Updated Date: 2024.12.16 01:00:00                  #
 # ================================================== #
 
 import datetime
@@ -24,6 +24,7 @@ from .indexing import Indexing
 from .llm import Llm
 from .chat import Chat
 from .metadata import Metadata
+from .ui import UI
 
 from .types.ctx import Ctx
 from .types.external import External
@@ -43,6 +44,7 @@ class Idx:
         self.storage = Storage(window)
         self.chat = Chat(window, self.storage)
         self.metadata = Metadata(window)
+        self.ui = UI(window)
 
         self.providers = {
             "json_file": JsonFileProvider(window),  # only for patching
@@ -406,7 +408,11 @@ class Idx:
             return self.items[store_id]
         return {}
 
-    def append(self, idx: str, files: Dict[str, str]):
+    def append(
+            self,
+            idx: str,
+            files: Dict[str, str]
+    ):
         """
         Append indexed files to index
 
@@ -454,7 +460,11 @@ class Idx:
                 self.items[store_id][idx].items[file_id]["id"] = doc_id
                 self.items[store_id][idx].items[file_id]["indexed_ts"] = ts
 
-    def remove_doc(self, idx: str, doc_id: str):
+    def remove_doc(
+            self,
+            idx: str,
+            doc_id: str
+    ):
         """
         Remove document from index
 
@@ -465,7 +475,11 @@ class Idx:
         if self.storage.remove_document(idx, doc_id):
             self.log("Removed document from index: " + idx + " - " + doc_id)
 
-    def remove_file(self, idx: str, file: str):
+    def remove_file(
+            self,
+            idx: str,
+            file: str
+    ):
         """
         Remove file from index
 
@@ -517,7 +531,7 @@ class Idx:
                     self.items[store_id][idx].id = idx
                     self.items[store_id][idx].name = idx
 
-    def get_idx_ids(self) -> List[str]:
+    def get_idx_ids(self) -> List[Dict[str, str]]:
         """
         Get list of indexes
 
@@ -527,7 +541,10 @@ class Idx:
         data =  self.window.core.config.get('llama.idx.list')
         if data is not None:
             for item in data:
-                ids.append(item['id'])
+                name = item['name']
+                if name is None or name == "":
+                    name = item['id']
+                ids.append({item['id']: name})
         return ids
 
     def clear(self, idx: str):

pygpt_net/core/idx/chat.py

@@ -6,7 +6,7 @@
 # GitHub:  https://github.com/szczyglis-dev/py-gpt   #
 # MIT License                                        #
 # Created By  : Marcin Szczygliński                  #
-# Updated Date: 2024.12.14 08:00:00                  #
+# Updated Date: 2024.12.14 22:00:00                  #
 # ================================================== #
 
 import json
@@ -360,7 +360,7 @@ class Chat:
             ctx: CtxItem,
             type: str,
             url: str,
-            args: dict,
+            args: Dict[str, Any],
             query: str,
             model: Optional[ModelItem] = None
     ) -> str:
@@ -552,7 +552,11 @@ class Chat:
         ]
         return ChatPromptTemplate(qa_msgs)
 
-    def get_index(self, idx: str, model: ModelItem):
+    def get_index(
+            self,
+            idx: str,
+            model: ModelItem
+    ):
         """
         Get index instance
 
@@ -573,7 +577,10 @@ class Chat:
         index = self.storage.get(idx, service_context=service_context)  # get index
         return index, service_context
 
-    def get_metadata(self, source_nodes: Optional[list]) -> Dict[str, Any]:
+    def get_metadata(
+            self,
+            source_nodes: Optional[list]
+    ) -> Dict[str, Any]:
         """
         Get metadata from source nodes
 
@@ -601,7 +608,10 @@ class Chat:
                             break
         return metadata
 
-    def parse_metadata(self, metadata: Optional[Dict]) -> Dict[str, Any]:
+    def parse_metadata(
+            self,
+            metadata: Optional[Dict]
+    ) -> Dict[str, Any]:
         """
         Parse metadata