downloadable classification, multiple files

app.py

@@ -5,6 +5,8 @@ from PyPDF2 import PdfReader
 from openai import OpenAI
 from langchain.chat_models import ChatOpenAI
 
+ASK_ASH_PASSWORD = os.environ["ASK_ASH_PASSWORD"]
+
 
 def gpt4_new(prompt_text):
     client = OpenAI(api_key=os.environ.get('OPENAI_API_KEY'))
@@ -80,47 +82,77 @@ def json_open(filename):
 
 
 def main():
-    ASK_ASH_PASSWORD = os.environ["ASK_ASH_PASSWORD"]
     st.title("Doc Classifier")
     st.subheader("Limitationen: ")
     st.write("bisher nur PDFs")
-    st.write("nur Disziplin und Doc typ.")
-    st.write("Dokumenten Grösse ist stark limitiert")
+    st.write("nur Disziplin, Doc typ. und Geschoss")
     st.write("macht noch viele Fehler, vor allem bei Koordination, Datennetz usw, (unklare Disziplinen)")
     st.write("")
 
     if st.text_input("ASK_ASH_PASSWORD: ", type="password") == ASK_ASH_PASSWORD:
-        uploaded_file = st.file_uploader("PDF Dokument", accept_multiple_files=False)
+        uploaded_files = st.file_uploader("PDF Dokument", accept_multiple_files=True)
         #print(uploaded_file)
         #print(uploaded_file.name)
+
         if st.button("classify KBOB!"):
-            if uploaded_file is not None:
-                with st.spinner("GPT4 at work"):
-                    pdf_text = str(get_pdf_text(uploaded_file))
-                    prompt_1 = auftrag_0 + auftrag_1_disz + str(Baubranchen_Disziplinen) + pdf_text
-                    answer_1 = gpt4_new(prompt_1)
-                    print(prompt_1)
-
-                st.write(answer_1)
-                with st.spinner("GPT4 at work"):
-                    prompt_2 = auftrag_0 + auftrag_1_dokt + str(Dokumententypen) + pdf_text
-                    answer_2 = gpt4_new(prompt_2)
-                    print(prompt_2)
-
-                st.write(answer_2)
-                with st.spinner("GPT4 at work"):
-                    prompt_3 = auftrag_0 + auftrag_1_gesch + str(ASH_Geschosse) + pdf_text
-                    answer_3 = gpt4_new(prompt_3)
-                    print(prompt_3)
-                    with open('gpt4_responses.txt', 'w', encoding='utf-8') as file:
-                        file.writelines(answer_1)
-                        file.writelines(answer_2)
-                        file.writelines(answer_3)
-                st.write(answer_3)
-                st.success("work done, saved")
-
-                st.download_button("download txt", json_open("gpt4_responses.txt"), file_name="classification.txt")
-                #st.download_button("download txt", json_open("gpt4_responses.txt"), file_name="gpt4_responses.txt")
+            if uploaded_files is not None:
+                with st.container():
+                    # col1, col2, col3, col4, col5 = st.columns(5)
+                    col1, col2, col3 = st.columns(3)
+                    all_metadata = []
+                    with col1:
+                        st.write("Disziplin")
+                        st.write(f"")
+                    with col2:
+                        st.write("Disziplin")
+                        st.write(f"")
+                    with col3:
+                        st.write("Disziplin")
+                        st.write(f"")
+
+                    for file in uploaded_files:
+                        metadata = [file.name]
+                        with col1:
+                            with st.spinner("GPT4 at work"):
+                                pdf_text = str(get_pdf_text(file))
+                                prompt_1 = auftrag_0 + auftrag_1_disz + str(Baubranchen_Disziplinen) + pdf_text
+                                answer_1 = gpt4_new(prompt_1)
+                                print(prompt_1)
+                                metadata.append(answer_1)
+                            st.write(answer_1)
+
+                        with col2:
+                            with st.spinner("GPT4 at work"):
+                                prompt_2 = auftrag_0 + auftrag_1_dokt + str(Dokumententypen) + pdf_text
+                                answer_2 = gpt4_new(prompt_2)
+                                print(prompt_2)
+                                metadata.append(answer_2)
+
+                            st.write(answer_2)
+
+                        with col3:
+                            with st.spinner("GPT4 at work"):
+                                prompt_3 = auftrag_0 + auftrag_1_gesch + str(ASH_Geschosse) + pdf_text
+                                answer_3 = gpt4_new(prompt_3)
+                                print(prompt_3)
+                                metadata.append(answer_2)
+
+                            st.write(answer_3)
+
+                        all_metadata.append(metadata)
+
+                    metadata_filename = "ai_generated_metadata.txt"
+                    with open(metadata_filename, 'w', encoding='utf-8') as f:
+                        for line in all_metadata:
+                            f.writelines("\n")
+                            for item in line:
+                                f.writelines(item)
+                                f.writelines(";")
+
+                            f.writelines("\n")
+
+                    st.success("classified, saved")
+                    st.download_button(f"Download Metadata", json_open(metadata_filename), file_name=metadata_filename)
             else:
                 st.warning("no file")