File size: 9,612 Bytes
bf6d237 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 |
"""This file should be imported only and only if you want to run the UI locally."""
import itertools
import logging
from collections.abc import Iterable
from pathlib import Path
from typing import Any
import gradio as gr # type: ignore
from fastapi import FastAPI
from gradio.themes.utils.colors import slate # type: ignore
from injector import inject, singleton
from llama_index.llms import ChatMessage, ChatResponse, MessageRole
from pydantic import BaseModel
from private_gpt.constants import PROJECT_ROOT_PATH
from private_gpt.di import global_injector
from private_gpt.server.chat.chat_service import ChatService, CompletionGen
from private_gpt.server.chunks.chunks_service import Chunk, ChunksService
from private_gpt.server.ingest.ingest_service import IngestService
from private_gpt.settings.settings import settings
from private_gpt.ui.images import logo_svg
logger = logging.getLogger(__name__)
THIS_DIRECTORY_RELATIVE = Path(__file__).parent.relative_to(PROJECT_ROOT_PATH)
# Should be "private_gpt/ui/avatar-bot.ico"
AVATAR_BOT = THIS_DIRECTORY_RELATIVE / "avatar-bot.ico"
UI_TAB_TITLE = "CHATBOT"
SOURCES_SEPARATOR = "\n\n Sources: \n"
class Source(BaseModel):
file: str
page: str
text: str
class Config:
frozen = True
@staticmethod
def curate_sources(sources: list[Chunk]) -> set["Source"]:
curated_sources = set()
for chunk in sources:
doc_metadata = chunk.document.doc_metadata
file_name = doc_metadata.get("file_name", "-") if doc_metadata else "-"
page_label = doc_metadata.get("page_label", "-") if doc_metadata else "-"
source = Source(file=file_name, page=page_label, text=chunk.text)
curated_sources.add(source)
return curated_sources
@singleton
class PrivateGptUi:
@inject
def __init__(
self,
ingest_service: IngestService,
chat_service: ChatService,
chunks_service: ChunksService,
) -> None:
self._ingest_service = ingest_service
self._chat_service = chat_service
self._chunks_service = chunks_service
# Cache the UI blocks
self._ui_block = None
def _chat(self, message: str, history: list[list[str]], mode: str, *_: Any) -> Any:
def yield_deltas(completion_gen: CompletionGen) -> Iterable[str]:
full_response: str = ""
stream = completion_gen.response
for delta in stream:
if isinstance(delta, str):
full_response += str(delta)
elif isinstance(delta, ChatResponse):
full_response += delta.delta or ""
yield full_response
if completion_gen.sources:
full_response += SOURCES_SEPARATOR
cur_sources = Source.curate_sources(completion_gen.sources)
sources_text = "\n\n\n".join(
f"{index}. {source.file} (page {source.page})"
for index, source in enumerate(cur_sources, start=1)
)
full_response += sources_text
yield full_response
def build_history() -> list[ChatMessage]:
history_messages: list[ChatMessage] = list(
itertools.chain(
*[
[
ChatMessage(content=interaction[0], role=MessageRole.USER),
ChatMessage(
# Remove from history content the Sources information
content=interaction[1].split(SOURCES_SEPARATOR)[0],
role=MessageRole.ASSISTANT,
),
]
for interaction in history
]
)
)
# max 20 messages to try to avoid context overflow
return history_messages[:20]
new_message = ChatMessage(content=message, role=MessageRole.USER)
all_messages = [*build_history(), new_message]
match mode:
case "Query Docs":
# Add a system message to force the behaviour of the LLM
# to answer only questions about the provided context.
all_messages.insert(
0,
ChatMessage(
content="You can only answer questions about the provided context. If you know the answer "
"but it is not based in the provided context, don't provide the answer, just state "
"the answer is not in the context provided.",
role=MessageRole.SYSTEM,
),
)
query_stream = self._chat_service.stream_chat(
messages=all_messages,
use_context=True,
)
yield from yield_deltas(query_stream)
case "LLM Chat":
llm_stream = self._chat_service.stream_chat(
messages=all_messages,
use_context=False,
)
yield from yield_deltas(llm_stream)
case "Search in Docs":
response = self._chunks_service.retrieve_relevant(
text=message, limit=4, prev_next_chunks=0
)
sources = Source.curate_sources(response)
yield "\n\n\n".join(
f"{index}. **{source.file} "
f"(page {source.page})**\n "
f"{source.text}"
for index, source in enumerate(sources, start=1)
)
def _list_ingested_files(self) -> list[list[str]]:
files = set()
for ingested_document in self._ingest_service.list_ingested():
if ingested_document.doc_metadata is None:
# Skipping documents without metadata
continue
file_name = ingested_document.doc_metadata.get(
"file_name", "[FILE NAME MISSING]"
)
files.add(file_name)
return [[row] for row in files]
def _upload_file(self, files: list[str]) -> None:
logger.debug("Loading count=%s files", len(files))
paths = [Path(file) for file in files]
self._ingest_service.bulk_ingest([(str(path.name), path) for path in paths])
def _build_ui_blocks(self) -> gr.Blocks:
logger.debug("Creating the UI blocks")
with gr.Blocks(
title=UI_TAB_TITLE,
theme=gr.themes.Soft(primary_hue=slate),
#css=".logo { "
#"display:flex;"
#"background-color: #C7BAFF;"
#"height: 80px;"
#"border-radius: 8px;"
#"align-content: center;"
#"justify-content: center;"
#"align-items: center;"
#"}"
#".logo img { height: 25% }",
) as blocks:
#with gr.Row():
#gr.HTML(f"<div class='logo'/><img src={logo_svg} alt=PrivateGPT></div")
with gr.Row():
with gr.Column(scale=3, variant="compact"):
mode = gr.Radio(
["Query Docs", "Search in Docs", "LLM Chat"],
label="Mode",
value="Query Docs",
)
upload_button = gr.components.UploadButton(
"Upload File(s)",
type="filepath",
file_count="multiple",
size="sm",
)
ingested_dataset = gr.List(
self._list_ingested_files,
headers=["File name"],
label="Ingested Files",
interactive=False,
render=False, # Rendered under the button
)
upload_button.upload(
self._upload_file,
inputs=upload_button,
outputs=ingested_dataset,
)
ingested_dataset.change(
self._list_ingested_files,
outputs=ingested_dataset,
)
ingested_dataset.render()
with gr.Column(scale=7):
_ = gr.ChatInterface(
self._chat,
chatbot=gr.Chatbot(
label=f"LLM: {settings().llm.mode}",
show_copy_button=True,
render=False,
avatar_images=(
None,
AVATAR_BOT,
),
),
additional_inputs=[mode, upload_button],
)
return blocks
def get_ui_blocks(self) -> gr.Blocks:
if self._ui_block is None:
self._ui_block = self._build_ui_blocks()
return self._ui_block
def mount_in_app(self, app: FastAPI, path: str) -> None:
blocks = self.get_ui_blocks()
blocks.queue()
logger.info("Mounting the gradio UI, at path=%s", path)
gr.mount_gradio_app(app, blocks, path=path)
if __name__ == "__main__":
ui = global_injector.get(PrivateGptUi)
_blocks = ui.get_ui_blocks()
_blocks.queue()
_blocks.launch(debug=False, show_api=False)
|