Initial commit — clean history (removed large test files, browser profiles, Medidata/Clario downloads)

2026-06-01 15:36:31 +02:00
commit bb604e593e
1304 changed files with 116480 additions and 0 deletions
@@ -0,0 +1,50 @@
+"""
+Import 77242113UCO3002_CZ_SK_Investigators_with_internet_summary.xlsx
+do MongoDB databáze 'feasibility', kolekce 'investigators'.
+
+Spustit: python import_to_mongo.py
+Závislosti: pip install openpyxl pymongo
+"""
+
+import openpyxl
+from pymongo import MongoClient
+import os
+
+XLSX = os.path.join(os.path.dirname(__file__),
+                    "77242113UCO3002_CZ_SK_Investigators_with_internet_summary.xlsx")
+MONGO_URI = "mongodb://192.168.1.76:27017/"
+DB_NAME = "feasibility"
+COLLECTION = "investigators"
+
+wb = openpyxl.load_workbook(XLSX, data_only=True)
+ws = wb["CZ + SK Investigators"]
+rows = list(ws.iter_rows(values_only=True))
+
+docs = []
+for row in rows[1:]:
+    if not any(row):
+        continue
+    zeme, prijmeni, jmeno, email, pracoviste, studie_raw, zdroje_raw, summary = row
+    docs.append({
+        "zeme":             zeme,
+        "prijmeni":         prijmeni,
+        "jmeno":            jmeno,
+        "email":            email,
+        "pracoviste":       pracoviste,
+        "studie":           [s.strip() for s in studie_raw.split("\n") if s.strip()] if studie_raw else [],
+        "zdroje":           [u.strip() for u in zdroje_raw.split("\n") if u.strip()] if zdroje_raw else [],
+        "internet_summary": summary,
+    })
+
+client = MongoClient(MONGO_URI)
+db = client[DB_NAME]
+col = db[COLLECTION]
+col.drop()                         # čistý start
+result = col.insert_many(docs)
+print(f"Vloženo {len(result.inserted_ids)} dokumentů do {DB_NAME}.{COLLECTION}")
+
+col.create_index("studie")
+col.create_index("zeme")
+col.create_index("prijmeni")
+print("Indexy vytvořeny: studie, zeme, prijmeni")
+client.close()