Upload 3 files

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
 earnings_calls_sentencewise.csv filter=lfs diff=lfs merge=lfs -text
+earnings_calls_cleaned_metadata.csv filter=lfs diff=lfs merge=lfs -text

app.py

@@ -72,7 +72,7 @@ with st.sidebar:
 
 # Choose encoder model
 
-encoder_models_choice = ["SGPT", "MPNET"]
+encoder_models_choice = ["MPNET", "SGPT"]
 with st.sidebar:
     encoder_model = st.selectbox("Select Encoder Model", encoder_models_choice)
 

earnings_calls_cleaned_metadata.csv

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6c6474da1f710d2a6d2ea65c475baf6821db95a5cb81dd8703eec3c04cd22cbe
+size 18988194

utils.py

@@ -17,7 +17,7 @@ import streamlit_scrollable_textbox as stx
 
 @st.experimental_singleton
 def get_data():
-    data = pd.read_csv("earnings_calls_sentencewise.csv")
+    data = pd.read_csv("earnings_calls_cleaned_metadata.csv")
     return data
 
 
@@ -72,6 +72,7 @@ def query_pinecone(query, top_k, model, index, year, quarter, ticker, threshold=
             "Year": int(year),
             "Quarter": {"$eq": quarter},
             "Ticker": {"$eq": ticker},
+            "QA_Flag": {"$eq": "Answer"},
         },
         include_metadata=True,
     )
@@ -103,7 +104,7 @@ def sentence_id_combine(data, query_results, lag=2):
     ]
     # Create a list of context sentences by joining the sentences corresponding to the lookup IDs
     context_list = [
-        ". ".join(data.Text.iloc[lookup_id].to_list()) for lookup_id in lookup_ids
+        " ".join(data.Text.iloc[lookup_id].to_list()) for lookup_id in lookup_ids
     ]
     return context_list
 
@@ -114,11 +115,11 @@ def text_lookup(data, sentence_ids):
 
 
 def generate_prompt(query_text, context_list):
-    #context = " ".join(context_list)
+    context = " \n".join(context_list)
     prompt = f"""
     Context information is below: 
     ---------------------
-    {context_list}
+    {context}
     ---------------------
     Given the context information and prior knowledge, answer this question:
     {query_text} 
@@ -148,15 +149,16 @@ def retrieve_transcript(data, year, quarter, ticker):
             (data.Year == int(year))
             & (data.Quarter == quarter)
             & (data.Ticker == ticker),
-            ["Year", "Month", "Date", "Ticker"],
+            ["File_Name"],
         ]
         .drop_duplicates()
-        .iloc[0]
+        .iloc[0,0]
     )
+    print(row)
     # convert row to a string and join values with "-"
-    row_str = "-".join(row.astype(str)) + ".txt"
+    #row_str = "-".join(row.astype(str)) + ".txt"
     open_file = open(
-        f"Transcripts/{ticker}/{row_str}",
+        f"Transcripts/{ticker}/{row}",
         "r",
     )
     file_text = open_file.read()