b7e656a319
New front-end Launch Chat API Manage Sources Enable re-embedding of all contents Sources can be added without a notebook now Improved settings Enable model selector on all chats Background processing for better experience Dark mode Improved Notes Improved Docs: - Remove all Streamlit references from documentation - Update deployment guides with React frontend setup - Fix Docker environment variables format (SURREAL_URL, SURREAL_PASSWORD) - Update docker image tag from :latest to :v1-latest - Change navigation references (Settings → Models to just Models) - Update development setup to include frontend npm commands - Add MIGRATION.md guide for users upgrading from Streamlit - Update quick-start guide with correct environment variables - Add port 5055 documentation for API access - Update project structure to reflect frontend/ directory - Remove outdated source-chat documentation files
115 lines
5.3 KiB
Python
115 lines
5.3 KiB
Python
import streamlit as st
|
|
from humanize import naturaltime
|
|
|
|
from api.insights_service import insights_service
|
|
from api.models_service import ModelsService
|
|
from api.sources_service import SourcesService
|
|
from api.transformations_service import TransformationsService
|
|
from pages.stream_app.utils import check_models
|
|
|
|
# Initialize service instances
|
|
sources_service = SourcesService()
|
|
transformations_service = TransformationsService()
|
|
models_service = ModelsService()
|
|
|
|
|
|
def source_panel(source_id: str, notebook_id=None, modal=False):
|
|
check_models(only_mandatory=False)
|
|
source_with_metadata = sources_service.get_source(source_id)
|
|
if not source_with_metadata:
|
|
raise ValueError(f"Source not found: {source_id}")
|
|
|
|
# Now we can access both the source and embedded_chunks directly
|
|
current_title = source_with_metadata.title if source_with_metadata.title else "No Title"
|
|
source_with_metadata.title = st.text_input("Title", value=current_title)
|
|
if source_with_metadata.title != current_title:
|
|
sources_service.update_source(source_with_metadata.source)
|
|
st.toast("Saved new Title")
|
|
|
|
process_tab, source_tab = st.tabs(["Process", "Source"])
|
|
with process_tab:
|
|
c1, c2 = st.columns([4, 2])
|
|
with c1:
|
|
title = st.empty()
|
|
if source_with_metadata.title:
|
|
title.subheader(source_with_metadata.title)
|
|
if source_with_metadata.asset and source_with_metadata.asset.url:
|
|
from_src = f"from URL: {source_with_metadata.asset.url}"
|
|
elif source_with_metadata.asset and source_with_metadata.asset.file_path:
|
|
from_src = f"from file: {source_with_metadata.asset.file_path}"
|
|
else:
|
|
from_src = "from text"
|
|
st.caption(f"Created {naturaltime(source_with_metadata.created)}, {from_src}")
|
|
for insight in insights_service.get_source_insights(source_with_metadata.id):
|
|
with st.expander(f"**{insight.insight_type}**"):
|
|
st.markdown(insight.content)
|
|
x1, x2 = st.columns(2)
|
|
if x1.button(
|
|
"Delete", type="primary", key=f"delete_insight_{insight.id}"
|
|
):
|
|
insights_service.delete_insight(insight.id or "")
|
|
st.rerun(scope="fragment" if modal else "app")
|
|
st.toast("Insight deleted")
|
|
if notebook_id:
|
|
if x2.button(
|
|
"Save as Note", icon="📝", key=f"save_note_{insight.id}"
|
|
):
|
|
insights_service.save_insight_as_note(insight.id or "", notebook_id)
|
|
st.toast("Saved as Note. Refresh the Notebook to see it.")
|
|
|
|
with c2:
|
|
transformations = transformations_service.get_all_transformations()
|
|
if transformations:
|
|
with st.container(border=True):
|
|
transformation = st.selectbox(
|
|
"Run a transformation",
|
|
transformations,
|
|
key=f"transformation_{source_with_metadata.id}",
|
|
format_func=lambda x: x.name,
|
|
)
|
|
st.caption(transformation.description if transformation else "")
|
|
if st.button("Run"):
|
|
insights_service.create_source_insight(
|
|
source_id=source_with_metadata.id,
|
|
transformation_id=transformation.id or ""
|
|
)
|
|
st.rerun(scope="fragment" if modal else "app")
|
|
else:
|
|
st.markdown(
|
|
"No transformations created yet. Create new Transformation to use this feature."
|
|
)
|
|
|
|
default_models = models_service.get_default_models()
|
|
embedding_model = default_models.default_embedding_model
|
|
if not embedding_model:
|
|
help = (
|
|
"No embedding model found. Please, select one on the Models page."
|
|
)
|
|
else:
|
|
help = "This will generate your embedding vectors on the database for powerful search capabilities"
|
|
|
|
if not source_with_metadata.embedded_chunks and st.button(
|
|
"Embed vectors",
|
|
icon="🦾",
|
|
help=help,
|
|
disabled=not embedding_model,
|
|
):
|
|
from api.embedding_service import embedding_service
|
|
|
|
result = embedding_service.embed_content(source_with_metadata.id, "source")
|
|
result_dict = result if isinstance(result, dict) else result[0] if isinstance(result, list) else {}
|
|
st.success(result_dict.get("message", "Embedding complete"))
|
|
|
|
with st.container(border=True):
|
|
st.caption(
|
|
"Deleting the source will also delete all its insights and embeddings"
|
|
)
|
|
if st.button(
|
|
"Delete", type="primary", key=f"bt_delete_source_{source_with_metadata.id}"
|
|
):
|
|
sources_service.delete_source(source_with_metadata.id)
|
|
st.rerun()
|
|
|
|
with source_tab:
|
|
st.subheader("Content")
|
|
st.markdown(source_with_metadata.full_text)
|