app.py

import json
import streamlit as st
from google.oauth2 import service_account
from google.cloud import language_v1
import urllib.parse
import urllib.request
import pandas as pd


# Function to query Google's Knowledge Graph API
def query_knowledge_graph(entity_id):
    try:
        google_search_link = f"https://www.google.com/search?kgmid={entity_id}"
        st.markdown(f'[Open in Google Search]({google_search_link})', unsafe_allow_html=True)
    except Exception as e:
        st.write(f"An error occurred: {e}")

# Function to count entities with 'mid' that contains '/g/' or '/m/' in their metadata
def count_entities(entities):
    count = 0
    for entity in entities:
        metadata = entity.metadata
        if 'mid' in metadata and ('/g/' in metadata['mid'] or '/m/' in metadata['mid']):
            count += 1
    return count

# Function to serialize entity metadata
def serialize_entity_metadata(metadata):
    return {k: str(v) for k, v in metadata.items()}

# Function to export entities as a JSON or CSV file, only exporting entities with 'mid' in their metadata
def export_entities(entities):
    entity_list = []
    for entity in entities:
        # Check if entity has 'mid' in its metadata and if it contains '/g/' or '/m/' in 'mid'
        if 'mid' in entity.metadata and ('/g/' in entity.metadata['mid'] or '/m/' in entity.metadata['mid']):
            entity_info = {
                "Name": entity.name,
                "Type": language_v1.Entity.Type(entity.type_).name,
                "Salience Score": entity.salience,
                "Metadata": serialize_entity_metadata(entity.metadata),
                "Mentions": [mention.text.content for mention in entity.mentions]
            }
            entity_list.append(entity_info)
    
    # If there are no entities to export, notify the user
    if not entity_list:
        st.write("No entities with a valid 'mid' found to export.")
        return
    
    # Convert to DataFrame for easier export as CSV
    df = pd.DataFrame(entity_list)
    
    # Export as CSV
    csv = df.to_csv(index=False)
    st.download_button(label="Export Entities as CSV", data=csv, file_name="entities.csv", mime="text/csv")
    
    # Export as JSON
    json_data = json.dumps(entity_list, indent=2)
    st.download_button(label="Export Entities as JSON", data=json_data, file_name="entities.json", mime="application/json")

# Sidebar content
st.sidebar.title("About This Tool")
st.sidebar.markdown("This tool leverages Google's NLP technology for entity analysis.")
st.sidebar.markdown("### Step-by-Step Guide")
st.sidebar.markdown("""
1. **Open the Tool**: Navigate to the URL where the tool is hosted.
2. **User Input**: Enter the text you want to analyze.
3. **Analyze**: Click the 'Analyze' button.
4. **View Results**: See the identified entities and their details.
5. **Export Entities**: Export the entities as JSON or CSV.
""")

# Header and intro
st.title("Google Cloud NLP Entity Analyzer")
st.write("This tool analyzes text to identify entities such as people, locations, organizations, and events.")
st.write("Entity salience scores are always relative to the analyzed text.")

def sample_analyze_entities(text_content):
    service_account_info = json.loads(st.secrets["google_nlp"])
    credentials = service_account.Credentials.from_service_account_info(
        service_account_info, scopes=["https://www.googleapis.com/auth/cloud-platform"]
    )
    
    client = language_v1.LanguageServiceClient(credentials=credentials)
    document = {"content": text_content, "type_": language_v1.Document.Type.PLAIN_TEXT, "language": "en"}
    encoding_type = language_v1.EncodingType.UTF8

    response = client.analyze_entities(request={"document": document, "encoding_type": encoding_type})
    
    # Count the entities with 'mid' and either '/g/' or '/m/' in their metadata
    entity_count = count_entities(response.entities)

    if entity_count == 0:
        st.markdown(f"# We found {len(response.entities)} entities - but found no Google Entities")
        st.write("---")
    elif entity_count == 1:
        st.markdown(f"# We found {len(response.entities)} entities - and found 1 Google Entity")
        st.write("---")
    else:
        st.markdown(f"# We found {len(response.entities)} entities - and found {entity_count} Google Entities")
        st.write("---")


    for i, entity in enumerate(response.entities):
        st.write(f"Entity {i+1} of {len(response.entities)}")
        st.write(f"Name: {entity.name}")
        st.write(f"Type: {language_v1.Entity.Type(entity.type_).name}")
        st.write(f"Salience Score: {entity.salience}")

        if entity.metadata:
            st.write("Metadata:")
            st.write(entity.metadata)
            
            if 'mid' in entity.metadata and ('/g/' in entity.metadata['mid'] or '/m/' in entity.metadata['mid']):
                entity_id = entity.metadata['mid']
                query_knowledge_graph(entity_id)

        if entity.mentions:
            mention_count = len(entity.mentions)
            plural = "s" if mention_count > 1 else ""
            st.write(f"Mentions: {mention_count} mention{plural}")
            st.write("Raw Array:")
            st.write(entity.mentions)

        st.write("---")

    # Add the export functionality
    export_entities(response.entities)

# User input for text analysis
user_input = st.text_area("Enter text to analyze")

if st.button("Analyze"):
    if user_input:
        sample_analyze_entities(user_input)