Update app.py

app.py

@@ -51,48 +51,10 @@ The response should be formatted as a structured JSON object, as shown below:
     // Additional measurements
   ]
 }
-### Example:
-If the patent discusses "A novel RF power amplifier operating at 2.4 GHz with a bandwidth of 20 MHz and an output power of 30 dBm," the output should be:
-{
-  "Patent_Title": "High-Efficiency RF Power Amplifier",
-  "Patent_Abstract": "A novel RF power amplifier with improved impedance matching for wireless communication devices.",
-  "Patent_Metadata": {
-    "Patent_Number": "US12345678B2",
-    "Filing_Date": "2024-06-20",
-    "Inventors": ["Jane Doe", "John Smith"],
-    "Assignee": "TechWave Inc."
-  },
-  "Content": [
-    {
-      "Measurement_substance": "RF power amplifier",
-      "Measured_value": "2.4",
-      "Measured_unit": "GHz",
-      "measurement_type": "operating frequency"
-    },
-    {
-      "Measurement_substance": "RF power amplifier",
-      "Measured_value": "20",
-      "Measured_unit": "MHz",
-      "measurement_type": "bandwidth"
-    },
-    {
-      "Measurement_substance": "RF power amplifier",
-      "Measured_value": "30",
-      "Measured_unit": "dBm",
-      "measurement_type": "output power"
-    }
-  ]
-}
-### Key Considerations:
-- Extract as much detailed information as possible based on the text.
-- Retain accuracy: Avoid inferring data not explicitly mentioned.
-- Follow the structured JSON format strictly for consistency.
-- Exclude any irrelevant or redundant information. Focus only on RF and related hardware technical data.
 """
 
-
 # Title and description
-st.title("Technical Measurements Extractor for Patents")
+st.title("Blah")
 st.write(
     "Analyze patents to extract physical measurements such as frequency, bandwidth, and more. "
     "Provide a date range to download patents and analyze them using GPT models."
@@ -113,10 +75,35 @@ model_choice = st.selectbox(
 
 logging_enabled = st.checkbox("Enable logging?", value=False, help="Toggle logging for debugging purposes.")
 
+# Keyword Management
+st.header("Manage Keywords")
+st.write("Add or delete keywords for filtering patents.")
+default_keywords = [
+    "RF", "Radio Frequency", "Wireless Communication", "Antenna", "Microwave", "Electromagnetic Waves",
+    "Beamforming", "5G", "6G", "Patch Antenna", "Dipole Antenna", "Phased Array", "Radiation Pattern", "IoT",
+    "Wireless Charging"
+]
+keywords_input = st.text_area(
+    "Enter keywords for filtering (comma-separated):", value=", ".join(default_keywords)
+)
+user_keywords = [kw.strip() for kw in keywords_input.split(",") if kw.strip()]
+
+# Field Selection
+st.header("Choose Fields for Filtering")
+fields = st.multiselect(
+    "Select fields to search for keywords:",
+    ["Title", "Abstract", "Claims", "Summary", "Detailed Description"],
+    default=["Title", "Abstract"]
+)
+
 # Run Analysis Button
 if st.button("Analyze Patents"):
     if not start_date_input or not end_date_input:
         st.error("Please enter both start and end dates!")
+    elif not user_keywords:
+        st.error("Please provide at least one keyword for filtering.")
+    elif not fields:
+        st.error("Please select at least one field for filtering.")
     else:
         try:
             # Parse date inputs
@@ -138,8 +125,18 @@ if st.button("Analyze Patents"):
                     st.stop()
                 st.success(f"{len(saved_patent_names)} patents found and processed!")
 
-            # Step 2: Analyze patents using GPT
-            random_patents = saved_patent_names[:num_patents_to_analyze]
+            # Step 2: Filter patents based on user input
+            with st.spinner("Filtering patents..."):
+                filtered_patents = preprocess_data.filter_rf_patents(
+                    saved_patent_names, keywords=user_keywords, fields=fields
+                )
+                if not filtered_patents:
+                    st.error("No patents matched the filtering criteria.")
+                    st.stop()
+                st.success(f"{len(filtered_patents)} relevant patents found and processed!")
+
+            # Step 3: Analyze patents using GPT
+            random_patents = filtered_patents[:num_patents_to_analyze]
             total_cost = 0
             results = []
 
@@ -147,18 +144,13 @@ if st.button("Analyze Patents"):
             for i, patent_file in enumerate(random_patents):
                 cost, output = qa_agent.call_QA_to_json(
                     PROMPT,
-                    start_date.year,  # Pass start_date year, month, and day
-                    start_date.month,
-                    start_date.day,
-                    saved_patent_names,
-                    i,
-                    logging_enabled,
-                    model_choice,
+                    start_date.year, start_date.month, start_date.day,
+                    saved_patent_names, i, logging_enabled, model_choice
                 )
                 total_cost += cost
                 results.append(output)
 
-            # Step 3: Display results
+            # Step 4: Display results
             st.write(f"**Total Cost:** ${total_cost:.4f}")
             st.write("### Analysis Results:")
             for idx, result in enumerate(results):
@@ -168,4 +160,4 @@ if st.button("Analyze Patents"):
         except ValueError as ve:
             st.error(f"Invalid date format: {ve}")
         except Exception as e:
-            st.error(f"An unexpected error occurred: {e}")
+            st.error(f"An unexpected error occurred: {e}")