rework wordparser

- add dropdown selector
- fix bug where Telefon: key got overwritten

src/logic/wordparser.py

@@ -4,7 +4,8 @@ from dataclasses import dataclass
 from src.backend import Semester
 from typing import Union, Any
 from src.logic.openai import name_tester, run_shortener, semester_converter
-
+import zipfile
+from bs4 import BeautifulSoup
 import loguru
 import sys
 
@@ -134,8 +135,11 @@ def word_docx_to_csv(path: str) -> list[pd.DataFrame]:
             row_data: list[Any] = []
             for cell in row.cells:
                 text = cell.text
+
                 text = text.replace("\n", "")
                 row_data.append(text)
+                if text == "Ihr Fach:":
+                    row_data.append(get_fach(path))
             data.append(row_data)
         df = pd.DataFrame(data)
         df.columns = df.iloc[0]
@@ -143,10 +147,27 @@ def word_docx_to_csv(path: str) -> list[pd.DataFrame]:
 
         m_data.append(df)
 
-
     return m_data
 
 
+def get_fach(path: str) -> str:
+    document = zipfile.ZipFile(path)
+    xml_data = document.read("word/document.xml")
+    document.close()
+
+    soup = BeautifulSoup(xml_data, "xml")
+    # text we need is in <w:p w14:paraId="12456A32" ... > -> w:r -> w:t
+    paragraphs = soup.find_all("w:p")
+    names = []
+    for para in paragraphs:
+        para_id = para.get("w14:paraId")
+        if para_id == "12456A32":
+            # get the data in the w:t
+            for run in para.find_all("w:r"):
+                data = run.find("w:t")
+                return data.contents[0]
+
+
 def makeDict():
     return {
         "work_author": None,
@@ -245,13 +266,12 @@ def word_to_semap(word_path: str) -> SemapDocument:
     log.info("Parsing Word Document {}", word_path)
     semap = SemapDocument()
     df = word_docx_to_csv(word_path)
-    print(df)
     apparatdata = df[0]
     apparatdata = apparatdata.to_dict()
-
     keys = list(apparatdata.keys())
+    print(apparatdata, keys)
 
-    appdata = {keys[i]: keys[i + 1] for i in range(0, len(keys), 2)}
+    appdata = {keys[i]: keys[i + 1] for i in range(0, len(keys) - 1, 2)}
     semap.phoneNumber = appdata["Telefon:"]
     semap.subject = appdata["Ihr Fach:"]
     semap.mail = appdata["Mailadresse:"]