XML to PSD

app.py

@@ -6,6 +6,9 @@ import logging
 from openai import OpenAI
 import json
 import textwrap
+import requests
+import xml.etree.ElementTree as ET
+from lxml import etree
 
 openAiClient = OpenAI(
     api_key=os.getenv("OPENAI_API_KEY")
@@ -52,7 +55,44 @@ def remove_markdown(text):
  
     return text.strip()
 
-# Fetch PCC-3 information
+
+def generate_xml_from_xsd(xsd_url, answers):
+    response = requests.get(xsd_url)
+    if response.status_code != 200:
+        raise Exception(f"Failed to fetch XSD file from {xsd_url}")
+    
+    xsd_root = etree.fromstring(response.content)    
+    schema = etree.XMLSchema(xsd_root)    
+    parser = etree.XMLParser(schema=schema)
+    root = ET.Element(xsd_root.find(".//xs:element", namespaces={"xs": "http://www.w3.org/2001/XMLSchema"}).get("name"))
+    
+    for field, value in answers.items():
+        element = ET.SubElement(root, field)
+        element.text = str(value)
+    
+    try:
+        schema.assertValid(root)
+    except etree.DocumentInvalid as e:
+        print(f"Generated XML is not valid according to the schema: {e}")
+    
+    xml_string = ET.tostring(root, encoding="unicode", method="xml")
+    
+    return xml_string
+
+def generate_pcc3_xml():
+    global pcc3_structure, answers
+    if pcc3_structure and 'xsd_file' in pcc3_structure:
+        try:
+            xml_string = generate_xml_from_xsd(pcc3_structure['xsd_file'], answers)
+            return xml_string
+        except Exception as e:
+            logger.error(f"Error generating XML: {e}")
+            return None
+    else:
+        logger.error("PCC3 structure or XSD file URL not available")
+        return None
+
+
 def fetch_pcc3_info(case_purpose):
     logger.info("Fetching PCC information to formulate questions")
     system_message = textwrap.dedent(f"""
@@ -66,6 +106,7 @@ def fetch_pcc3_info(case_purpose):
     {{
         "form_description": "A brief description of the PCC form",
         "xsd_file": "Direct link to the concrete XSD file that was used to create the questions. This link should not point to anything else than XSD file. ",
+        "xsd_file_content": "Content of the XSD file. This should be a string that contains the content of the XSD file used to create the questions.",
         "form_id": "ID of the form that should be used to fill in the answers",
         "required_fields": [
             {{
@@ -77,7 +118,7 @@ def fetch_pcc3_info(case_purpose):
         ]
     }}
     Odpowiedz tylko JSON'em i niczym więcej. Nie stosuj zadnego formatowania. 
-    Nie stosuj zadnego wstępu. Wybierz najnowszy plik xsd i uzyj go do stworzenia pytan do uzytkownika. 
+    Nie stosuj zadnego wstępu. Wybierz plik xsd ktory najbardziej pasuje do podanego przez uzytkownika zadania i uzyj go do stworzenia pytan do uzytkownika. 
     Pytania z formularza powinny miec takie same nazwy pola jak pola w pliku xsd. 
     Uzyj nastepujacej strony w celu znalezienia odpowiedniego formularza do wypelnienia: https://www.podatki.gov.pl/pcc-sd/e-deklaracje-pcc-sd/formularze-pcc/
     """)

requirements.txt

@@ -2,4 +2,4 @@ gradio==4.44.0
 openai==1.50.2
 requests==2.32.3
 beautifulsoup4==4.12.3
-git+https://github.com/nathanrchn/perplexityai.git
+lxml==4.9.3