Spaces:
Build error
Build error
Update utils/legal_notebook_interface.py
Browse files
utils/legal_notebook_interface.py
CHANGED
|
@@ -2,13 +2,14 @@ import streamlit as st
|
|
| 2 |
from typing import List, Dict
|
| 3 |
import datetime
|
| 4 |
|
|
|
|
| 5 |
class LegalNotebookInterface:
|
| 6 |
def __init__(self, case_manager, vector_store, document_processor):
|
| 7 |
"""Initialize the notebook interface with required components."""
|
| 8 |
self.case_manager = case_manager
|
| 9 |
self.vector_store = vector_store
|
| 10 |
self.document_processor = document_processor
|
| 11 |
-
|
| 12 |
# Initialize session state
|
| 13 |
if "active_documents" not in st.session_state:
|
| 14 |
st.session_state.active_documents = []
|
|
@@ -20,37 +21,40 @@ class LegalNotebookInterface:
|
|
| 20 |
def render(self):
|
| 21 |
"""Render the main notebook interface."""
|
| 22 |
# Add custom styling
|
| 23 |
-
st.markdown(
|
| 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 |
# Create two-column layout
|
| 56 |
col1, col2 = st.columns([2, 3])
|
|
@@ -66,7 +70,7 @@ class LegalNotebookInterface:
|
|
| 66 |
def _render_document_panel(self):
|
| 67 |
"""Render the document viewing and management panel."""
|
| 68 |
st.markdown("### 📚 Documents")
|
| 69 |
-
|
| 70 |
# Document selection
|
| 71 |
cases = self.case_manager.get_all_cases()
|
| 72 |
if not cases:
|
|
@@ -78,11 +82,11 @@ class LegalNotebookInterface:
|
|
| 78 |
cases,
|
| 79 |
format_func=lambda x: x['title']
|
| 80 |
)
|
| 81 |
-
|
| 82 |
if selected_case:
|
| 83 |
st.session_state.current_case = selected_case['id']
|
| 84 |
documents = self.case_manager.list_documents(selected_case['id'])
|
| 85 |
-
|
| 86 |
if documents:
|
| 87 |
st.write("#### Available Documents")
|
| 88 |
for doc in documents:
|
|
@@ -97,7 +101,7 @@ class LegalNotebookInterface:
|
|
| 97 |
def _render_chat_panel(self):
|
| 98 |
"""Render the chat and analysis interface."""
|
| 99 |
st.markdown("### 💬 Document Analysis")
|
| 100 |
-
|
| 101 |
if not st.session_state.current_case:
|
| 102 |
st.info("Please select a case from the left panel to start analysis.")
|
| 103 |
return
|
|
@@ -105,49 +109,51 @@ class LegalNotebookInterface:
|
|
| 105 |
# Display chat history
|
| 106 |
for message in st.session_state.chat_history:
|
| 107 |
message_class = "user-message" if message["role"] == "user" else "assistant-message"
|
| 108 |
-
st.markdown(
|
|
|
|
| 109 |
<div class="message {message_class}">
|
| 110 |
{message["content"]}
|
| 111 |
</div>
|
| 112 |
-
|
|
|
|
|
|
|
| 113 |
|
| 114 |
# Chat input
|
| 115 |
if prompt := st.chat_input("Ask about your documents..."):
|
| 116 |
self._handle_chat_input(prompt)
|
| 117 |
|
| 118 |
def _handle_chat_input(self, prompt: str):
|
| 119 |
-
|
| 120 |
-
|
| 121 |
-
|
| 122 |
-
return
|
| 123 |
-
|
| 124 |
-
# Add user message
|
| 125 |
-
st.session_state.chat_history.append({
|
| 126 |
-
"role": "user",
|
| 127 |
-
"content": prompt,
|
| 128 |
-
"timestamp": datetime.datetime.now().isoformat()
|
| 129 |
-
})
|
| 130 |
-
|
| 131 |
-
try:
|
| 132 |
-
# Get documents from current case
|
| 133 |
-
documents = self.case_manager.list_documents(st.session_state.current_case)
|
| 134 |
-
if not documents:
|
| 135 |
-
st.error("No documents available for analysis.")
|
| 136 |
return
|
| 137 |
|
| 138 |
-
#
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 139 |
try:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 140 |
results = self.vector_store.similarity_search(
|
| 141 |
query=prompt,
|
| 142 |
k=3 # Number of chunks to retrieve
|
| 143 |
)
|
| 144 |
-
|
| 145 |
if not results:
|
| 146 |
response = "I couldn't find relevant information in the documents. Please try rephrasing your question."
|
| 147 |
else:
|
| 148 |
# Generate response from results
|
| 149 |
response = self._generate_response(prompt, results)
|
| 150 |
-
|
| 151 |
# Add response to chat history
|
| 152 |
st.session_state.chat_history.append({
|
| 153 |
"role": "assistant",
|
|
@@ -163,39 +169,36 @@ class LegalNotebookInterface:
|
|
| 163 |
"timestamp": datetime.datetime.now().isoformat()
|
| 164 |
})
|
| 165 |
|
| 166 |
-
except Exception as e:
|
| 167 |
-
st.error(f"Error analyzing documents: {str(e)}")
|
| 168 |
-
|
| 169 |
def _generate_response(self, prompt: str, results: List[Dict]) -> str:
|
| 170 |
-
|
| 171 |
-
|
| 172 |
-
|
| 173 |
-
|
| 174 |
-
|
| 175 |
-
|
| 176 |
-
|
| 177 |
-
|
| 178 |
-
|
| 179 |
-
|
| 180 |
-
|
| 181 |
-
|
| 182 |
-
|
| 183 |
-
|
| 184 |
-
|
| 185 |
-
|
| 186 |
|
| 187 |
def _show_document_details(self, doc: Dict):
|
| 188 |
"""Show detailed document information."""
|
| 189 |
st.markdown("#### Document Details")
|
| 190 |
st.write("**Title:** ", doc['title'])
|
| 191 |
st.write("**Added:** ", doc.get('added_at', 'Unknown'))
|
| 192 |
-
|
| 193 |
metadata = doc.get('metadata', {})
|
| 194 |
if metadata:
|
| 195 |
st.write("**Document Type:** ", metadata.get('document_type', 'Unknown'))
|
| 196 |
st.write("**Word Count:** ", metadata.get('word_count', 'Unknown'))
|
| 197 |
-
|
| 198 |
if 'entities' in metadata:
|
| 199 |
st.write("**Named Entities:** ")
|
| 200 |
for entity_type, entities in metadata['entities'].items():
|
| 201 |
-
st.write(f"- {entity_type}: {', '.join(entities[:5])}")
|
|
|
|
| 2 |
from typing import List, Dict
|
| 3 |
import datetime
|
| 4 |
|
| 5 |
+
|
| 6 |
class LegalNotebookInterface:
|
| 7 |
def __init__(self, case_manager, vector_store, document_processor):
|
| 8 |
"""Initialize the notebook interface with required components."""
|
| 9 |
self.case_manager = case_manager
|
| 10 |
self.vector_store = vector_store
|
| 11 |
self.document_processor = document_processor
|
| 12 |
+
|
| 13 |
# Initialize session state
|
| 14 |
if "active_documents" not in st.session_state:
|
| 15 |
st.session_state.active_documents = []
|
|
|
|
| 21 |
def render(self):
|
| 22 |
"""Render the main notebook interface."""
|
| 23 |
# Add custom styling
|
| 24 |
+
st.markdown(
|
| 25 |
+
"""
|
| 26 |
+
<style>
|
| 27 |
+
.document-panel {
|
| 28 |
+
background-color: #f8f9fa;
|
| 29 |
+
border-radius: 10px;
|
| 30 |
+
padding: 20px;
|
| 31 |
+
margin: 10px;
|
| 32 |
+
}
|
| 33 |
+
.message {
|
| 34 |
+
padding: 15px;
|
| 35 |
+
margin: 10px 0;
|
| 36 |
+
border-radius: 8px;
|
| 37 |
+
}
|
| 38 |
+
.user-message {
|
| 39 |
+
background-color: #f0f7ff;
|
| 40 |
+
border-left: 4px solid #1976d2;
|
| 41 |
+
}
|
| 42 |
+
.assistant-message {
|
| 43 |
+
background-color: #f8f9fa;
|
| 44 |
+
border-left: 4px solid #4caf50;
|
| 45 |
+
}
|
| 46 |
+
.suggestion-chip {
|
| 47 |
+
display: inline-block;
|
| 48 |
+
padding: 5px 15px;
|
| 49 |
+
margin: 5px;
|
| 50 |
+
background-color: #e3f2fd;
|
| 51 |
+
border-radius: 20px;
|
| 52 |
+
cursor: pointer;
|
| 53 |
+
}
|
| 54 |
+
</style>
|
| 55 |
+
""",
|
| 56 |
+
unsafe_allow_html=True,
|
| 57 |
+
)
|
| 58 |
|
| 59 |
# Create two-column layout
|
| 60 |
col1, col2 = st.columns([2, 3])
|
|
|
|
| 70 |
def _render_document_panel(self):
|
| 71 |
"""Render the document viewing and management panel."""
|
| 72 |
st.markdown("### 📚 Documents")
|
| 73 |
+
|
| 74 |
# Document selection
|
| 75 |
cases = self.case_manager.get_all_cases()
|
| 76 |
if not cases:
|
|
|
|
| 82 |
cases,
|
| 83 |
format_func=lambda x: x['title']
|
| 84 |
)
|
| 85 |
+
|
| 86 |
if selected_case:
|
| 87 |
st.session_state.current_case = selected_case['id']
|
| 88 |
documents = self.case_manager.list_documents(selected_case['id'])
|
| 89 |
+
|
| 90 |
if documents:
|
| 91 |
st.write("#### Available Documents")
|
| 92 |
for doc in documents:
|
|
|
|
| 101 |
def _render_chat_panel(self):
|
| 102 |
"""Render the chat and analysis interface."""
|
| 103 |
st.markdown("### 💬 Document Analysis")
|
| 104 |
+
|
| 105 |
if not st.session_state.current_case:
|
| 106 |
st.info("Please select a case from the left panel to start analysis.")
|
| 107 |
return
|
|
|
|
| 109 |
# Display chat history
|
| 110 |
for message in st.session_state.chat_history:
|
| 111 |
message_class = "user-message" if message["role"] == "user" else "assistant-message"
|
| 112 |
+
st.markdown(
|
| 113 |
+
f"""
|
| 114 |
<div class="message {message_class}">
|
| 115 |
{message["content"]}
|
| 116 |
</div>
|
| 117 |
+
""",
|
| 118 |
+
unsafe_allow_html=True,
|
| 119 |
+
)
|
| 120 |
|
| 121 |
# Chat input
|
| 122 |
if prompt := st.chat_input("Ask about your documents..."):
|
| 123 |
self._handle_chat_input(prompt)
|
| 124 |
|
| 125 |
def _handle_chat_input(self, prompt: str):
|
| 126 |
+
"""Process user input and generate response."""
|
| 127 |
+
if not st.session_state.current_case:
|
| 128 |
+
st.error("Please select a case first.")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 129 |
return
|
| 130 |
|
| 131 |
+
# Add user message
|
| 132 |
+
st.session_state.chat_history.append({
|
| 133 |
+
"role": "user",
|
| 134 |
+
"content": prompt,
|
| 135 |
+
"timestamp": datetime.datetime.now().isoformat()
|
| 136 |
+
})
|
| 137 |
+
|
| 138 |
try:
|
| 139 |
+
# Get documents from current case
|
| 140 |
+
documents = self.case_manager.list_documents(st.session_state.current_case)
|
| 141 |
+
if not documents:
|
| 142 |
+
st.error("No documents available for analysis.")
|
| 143 |
+
return
|
| 144 |
+
|
| 145 |
+
# Get relevant chunks
|
| 146 |
results = self.vector_store.similarity_search(
|
| 147 |
query=prompt,
|
| 148 |
k=3 # Number of chunks to retrieve
|
| 149 |
)
|
| 150 |
+
|
| 151 |
if not results:
|
| 152 |
response = "I couldn't find relevant information in the documents. Please try rephrasing your question."
|
| 153 |
else:
|
| 154 |
# Generate response from results
|
| 155 |
response = self._generate_response(prompt, results)
|
| 156 |
+
|
| 157 |
# Add response to chat history
|
| 158 |
st.session_state.chat_history.append({
|
| 159 |
"role": "assistant",
|
|
|
|
| 169 |
"timestamp": datetime.datetime.now().isoformat()
|
| 170 |
})
|
| 171 |
|
|
|
|
|
|
|
|
|
|
| 172 |
def _generate_response(self, prompt: str, results: List[Dict]) -> str:
|
| 173 |
+
"""Generate response based on search results."""
|
| 174 |
+
if not results:
|
| 175 |
+
return "I couldn't find relevant information in the documents."
|
| 176 |
+
|
| 177 |
+
# Format the response with chunks
|
| 178 |
+
response = "Based on the documents, here's what I found:\n\n"
|
| 179 |
+
|
| 180 |
+
for i, result in enumerate(results, 1):
|
| 181 |
+
response += f"**Excerpt {i}**\n"
|
| 182 |
+
response += f"{result.get('text', '')[:200]}...\n\n"
|
| 183 |
+
if result.get('metadata', {}).get('title'):
|
| 184 |
+
response += f"*Source: {result['metadata']['title']}*\n\n"
|
| 185 |
+
|
| 186 |
+
response += "\nWould you like me to elaborate on any of these points?"
|
| 187 |
+
|
| 188 |
+
return response
|
| 189 |
|
| 190 |
def _show_document_details(self, doc: Dict):
|
| 191 |
"""Show detailed document information."""
|
| 192 |
st.markdown("#### Document Details")
|
| 193 |
st.write("**Title:** ", doc['title'])
|
| 194 |
st.write("**Added:** ", doc.get('added_at', 'Unknown'))
|
| 195 |
+
|
| 196 |
metadata = doc.get('metadata', {})
|
| 197 |
if metadata:
|
| 198 |
st.write("**Document Type:** ", metadata.get('document_type', 'Unknown'))
|
| 199 |
st.write("**Word Count:** ", metadata.get('word_count', 'Unknown'))
|
| 200 |
+
|
| 201 |
if 'entities' in metadata:
|
| 202 |
st.write("**Named Entities:** ")
|
| 203 |
for entity_type, entities in metadata['entities'].items():
|
| 204 |
+
st.write(f"- {entity_type}: {', '.join(entities[:5])}")
|