Add application file

app.py

@@ -0,0 +1,113 @@
+import base64
+import hashlib
+import requests
+import os
+import streamlit as st
+
+# BOOKSHINE_HOST = 'https://api.bookshine.at'
+BOOKSHINE_HOST = 'http://localhost'
+BOOKSHINE_PORT = '4690'
+PASSWORDS = os.environ.get("PASSWORDS", "").split(",")
+
+def password_is_correct(password):
+    return password in PASSWORDS
+
+def process_docx(uploaded_file):
+    # Read and encode file contents
+    encoded_file = base64.b64encode(uploaded_file.read()).decode('utf-8')
+
+    payload = {
+        "filename": uploaded_file.name,
+        "attachment_b64": encoded_file,
+        "include_diff": True,
+        "include_docx": True,
+        "include_commas": True,
+        "is_demo": False,
+    }
+
+    response = requests.post(f'{BOOKSHINE_HOST}:{BOOKSHINE_PORT}/process_md', json=payload)
+
+    # Decode response and create download link
+    if response.ok:
+        response = response.json()
+
+        # make corrected docx available for download
+        decoded_file = base64.b64decode(response['docx'])
+        out_name = uploaded_file.name.replace('.docx', '-bookshine.docx')
+        st.download_button('✅ Korrigierten Umbruch herunterladen', data=decoded_file, file_name=out_name)
+
+        # show diff
+        diff = response["diff"].replace("¶", " ")
+        diff = diff.replace("<ins>", "<ins style='background-color: #c8e6c9; text-decoration: underline;'>")
+        st.write(diff, unsafe_allow_html=True)
+    else:
+        st.error('Es ist ein Fehler aufgetreten')
+        return
+
+
+def process_pdf(uploaded_file):
+    # Read and encode file contents
+    encoded_file = base64.b64encode(uploaded_file.read()).decode('utf-8')
+
+    # md5 encode file contents
+    md5 = hashlib.md5(encoded_file.encode()).hexdigest()
+
+    payload = {
+        "attachment_id": md5,
+        "attachment_b64": encoded_file,
+    }
+
+    response = requests.post(f'{BOOKSHINE_HOST}:{BOOKSHINE_PORT}/process_umbruch', json=payload)
+
+    # Decode response and create download link
+    if response.ok:
+        response = response.json()
+        # trennungen = [a for a in response["annotations"] if a["type"] == "TRENNUNG"]
+        b64 = response['annotated']
+        decoded_file = base64.b64decode(b64)
+        out_name = uploaded_file.name.replace('.pdf', '-bookshine.pdf')
+        st.download_button('✅ Report herunterladen', data=decoded_file, file_name=out_name)
+
+        # show pdf
+        pdf_display = f'<iframe src="data:application/pdf;base64,{b64}" width="800" height="800" type="application/pdf"></iframe>'
+        st.write(pdf_display, unsafe_allow_html=True)
+    else:
+        st.error('Es ist ein Fehler aufgetreten')
+
+
+def main():
+    st.title('Bookshine')
+
+    if "in_progress" not in st.session_state:
+        st.session_state.in_progress = False
+
+    # st.write("Bitte geben Sie die E-Mail-Adresse an, an die Sie die Resultate geschickt haben wollen.")
+    st.write("Wenn Sie Bookshine testen wollen, aber keinen Zugriffscode haben, schreiben Sie bitte eine E-Mail an [email protected].")
+
+    password = st.text_input("Zugriffscode")
+
+    uploaded_file = st.file_uploader('Manuskript (.docx) oder Umbruch (.pdf) hochladen', type=['docx', 'pdf'], accept_multiple_files=False)
+
+    if not uploaded_file:
+        return
+
+    process_button_name = "✨ Manuskript korrigieren"
+    if uploaded_file.name.endswith('.pdf'):
+        process_button_name = "✨ Umbruch prüfen"
+
+    def set_in_progress():
+        st.session_state.in_progress = True
+
+    # API request button
+    if st.button(process_button_name, on_click=set_in_progress, disabled=not password_is_correct(password) or st.session_state.in_progress):
+        with st.spinner('Datei wird verarbeitet …'):
+            if uploaded_file.name.endswith('.docx'):
+                process_docx(uploaded_file)
+            elif uploaded_file.name.endswith('.pdf'):
+                process_pdf(uploaded_file)
+            else:
+                st.warning('Bitte laden Sie eine Datei des Typs .docx oder .pdf hoch.')
+        st.session_state.in_progress = False
+
+if __name__ == '__main__':
+    main()

requirements.txt

@@ -0,0 +1,2 @@
+requests
+streamlit