Initial commit: What To Cook (WTC) v4.4 - Digitales Kochbuch, Wochenmenueplan & Buchdruck

summarize_texts.py

@@ -0,0 +1,17 @@
+import os
+import json
+
+TXT_DIR = '/home/maxi/Projekt/wtc.waiswuaschd.de/data/extracted_texts'
+files = sorted(os.listdir(TXT_DIR))
+
+summary = {}
+for f in files:
+    path = os.path.join(TXT_DIR, f)
+    with open(path, 'r', encoding='utf-8', errors='ignore') as fp:
+        content = fp.read()
+    summary[f] = content
+
+with open('/home/maxi/Projekt/wtc.waiswuaschd.de/data/recipes_summary.json', 'w', encoding='utf-8') as fp:
+    json.dump(summary, fp, ensure_ascii=False, indent=2)
+
+print("Saved recipes_summary.json with", len(summary), "files.")