notebook

2026-06-14 08:25:15 +02:00
parent f94573ea6e
commit ed6455787a
7 changed files with 876 additions and 20 deletions
@@ -0,0 +1,59 @@
+#!/usr/bin/env python3
+# -*- coding: utf-8 -*-
+"""Ktere zpravy v dane slozce chybi v Mongo (online vs archiv) — charakteristika."""
+import sys, collections
+import msal, requests
+from pymongo import MongoClient
+
+TENANT="7d269944-37a4-43a1-8140-c7517dc426e9"
+CLIENT="4b222bfd-78c9-4239-a53f-43006b3ed07f"
+SECRET="Txg8Q~MjhocuopxsJyJBhPmDfMxZ2r5WpTFj1dfk"
+GRAPH="https://graph.microsoft.com/v1.0"
+MBOX="michaela.buzalkova@buzalka.cz"
+TARGETS=["Inbox","Odloženo","Deleted Items","Sent Items"]  # displayName
+
+app=msal.ConfidentialClientApplication(CLIENT,authority=f"https://login.microsoftonline.com/{TENANT}",client_credential=SECRET)
+tok=app.acquire_token_for_client(scopes=["https://graph.microsoft.com/.default"])["access_token"]
+H={"Authorization":f"Bearer {tok}"}
+
+def allfolders(parent=None,ppath=""):
+    url=f"{GRAPH}/users/{MBOX}/mailFolders" if parent is None else f"{GRAPH}/users/{MBOX}/mailFolders/{parent}/childFolders"
+    params={"$top":100,"$select":"id,displayName,childFolderCount"}; out=[]
+    while url:
+        d=requests.get(url,headers=H,params=params,timeout=60).json()
+        for f in d.get("value",[]):
+            path=f"{ppath}/{f['displayName']}".lstrip("/")
+            out.append((path,f["displayName"],f["id"]))
+            if f.get("childFolderCount",0)>0: out+=allfolders(f["id"],path)
+        url=d.get("@odata.nextLink"); params=None
+    return out
+
+folders=allfolders()
+cli=MongoClient("mongodb://192.168.1.76:27017"); col=cli["emaily"][MBOX]
+
+for path,dname,fid in folders:
+    if dname not in TARGETS: continue
+    # online IDs
+    url=f"{GRAPH}/users/{MBOX}/mailFolders/{fid}/messages"
+    params={"$select":"internetMessageId,receivedDateTime,sentDateTime,from,subject","$top":100}
+    online={}
+    while url:
+        d=requests.get(url,headers=H,params=params,timeout=60).json()
+        for m in d.get("value",[]):
+            mid=m.get("internetMessageId")
+            if mid: online[mid]=m
+        url=d.get("@odata.nextLink"); params=None
+    have={x["_id"] for x in col.find({"folder_path":path},{"_id":1})}
+    missing=[m for mid,m in online.items() if mid not in have]
+    print(f"\n=== {path}  (online {len(online)}, mongo {len(have)}, chybi {len(missing)}) ===")
+    if not missing: continue
+    dts=sorted([(m.get("receivedDateTime") or m.get("sentDateTime") or "")[:10] for m in missing])
+    print(f"  datumove rozpeti chybejicich: {dts[0]} .. {dts[-1]}")
+    by=collections.Counter(((m.get('from') or {}).get('emailAddress') or {}).get('address','?') for m in missing)
+    print("  top odesilatele chybejicich:")
+    for a,n in by.most_common(8): print(f"     {n:>4}  {a}")
+    print("  ukazka 5 nejnovejsich:")
+    for m in sorted(missing,key=lambda x:(x.get('receivedDateTime') or ''),reverse=True)[:5]:
+        d=(m.get('receivedDateTime') or m.get('sentDateTime') or '')[:16]
+        s=(m.get('subject') or '')[:55]
+        print(f"     {d}  {s}")