fix: handling of dates

Author: oesteban

src/api.py

@@ -92,29 +92,33 @@ def fetch_window(
         if cursor:
             url = base_url + f"&cursor={cursor}"
 
-        r = requests.get(url, headers=headers)
-        if r.status_code == 429:
+        req = requests.get(url, headers=headers)
+        if req.status_code == 429:
             # parse Retry‑After or use exponential backoff
-            wait = int(r.headers.get("Retry-After", 2**errors))
+            wait = int(req.headers.get("Retry-After", 2**errors))
             sleep(wait + 0.1)
             errors += 1
             if errors > max_errors:
                 print("")
                 print(f"[{event_name}][{window_start}] too many 429s; abort")
                 break
             continue
-        elif not r.ok:
+        elif not req.ok:
             errors += 1
             if errors > max_errors:
                 print("")
-                print(f"[{event_name}][{window_start}] errors: {r.status_code}; abort")
+                print(
+                    f"[{event_name}][{window_start}] errors: {req.status_code}; abort"
+                )
                 break
             sleep(errors)  # simple backoff
             continue
 
         errors = 0
         events = [
-            e for e in r.json() if {"key": "environment", "value": "prod"} in e["tags"]
+            e
+            for e in req.json()
+            if {"key": "environment", "value": "prod"} in e["tags"]
         ]
         new_docs = [filter_new(e, db) for e in events]
         new_docs = [d for d in new_docs if d]
@@ -131,7 +135,7 @@ def fetch_window(
                 break
 
         # look at Link header for next cursor or end
-        link_header = r.headers.get("Link", "")
+        link_header = req.headers.get("Link", "")
         next_link = None
         if link_header:
             for link in parse_header_links(link_header):
@@ -165,10 +169,10 @@ def parallel_fetch(
         max_workers = os.cpu_count()
 
     windows = []
-    cur = since
-    while cur < until:
-        nxt = min(cur + datetime.timedelta(days=days_per_chunk), until)
-        windows.append((cur, nxt))
+    cur = until
+    while cur > since:
+        nxt = max(cur - datetime.timedelta(days=days_per_chunk), since)
+        windows.append((nxt, cur))
         cur = nxt
 
     kwargs = {

src/run.py

@@ -49,19 +49,26 @@ def cli():
     help="Which Sentry issues to fetch",
 )
 @click.option(
-    "-s",
-    "--since",
+    "-S",
+    "--start-date",
     type=click.DateTime(formats=["%Y-%m-%d"]),
     default=None,
     help=f"Start date (inclusive) in YYYY-MM-DD; defaults to {DEFAULT_DAYS_WINDOW} days ago",
 )
 @click.option(
-    "-u",
-    "--until",
+    "-E",
+    "--end-date",
     type=click.DateTime(formats=["%Y-%m-%d"]),
     default=None,
     help="End date (exclusive) in YYYY-MM-DD; defaults to today",
 )
+@click.option(
+    "-D",
+    "--days",
+    type=int,
+    default=None,
+    help="End date (exclusive) in YYYY-MM-DD; defaults to today",
+)
 @click.option(
     "-c",
     "--chunk-days",
@@ -80,7 +87,7 @@ def cli():
     "-M", "--max-errors", type=click.IntRange(min=1), default=DEFAULT_MAX_ERRORS
 )
 @click.option("-L", "--cached-limit", type=click.IntRange(min=1), default=None)
-def get(event, since, until, chunk_days, jobs, max_errors, cached_limit):
+def get(event, start_date, end_date, days, chunk_days, jobs, max_errors, cached_limit):
     """Fetch events in parallel using time-window chunking."""
 
     token = os.getenv("SENTRY_TOKEN")
@@ -90,18 +97,50 @@ def get(event, since, until, chunk_days, jobs, max_errors, cached_limit):
 
     now = datetime.now(timezone.utc)
 
-    since = since or (now - timedelta(days=DEFAULT_DAYS_WINDOW))
-    until = until or now
+    if start_date and days:
+        click.echo("Warning: overriding --start-date because --days was present")
+        start_date = None
+
+    if start_date:
+        start_date = datetime(
+            year=start_date.year,
+            month=start_date.month,
+            day=start_date.day,
+            hour=now.hour,
+            minute=now.minute,
+            second=now.second,
+            microsecond=now.microsecond,
+            tzinfo=now.tzinfo,
+        )
+    if end_date:
+        end_date = datetime(
+            year=end_date.year,
+            month=end_date.month,
+            day=end_date.day,
+            hour=now.hour,
+            minute=now.minute,
+            second=now.second,
+            microsecond=now.microsecond,
+            tzinfo=now.tzinfo,
+        )
+    else:
+        end_date = now
+
+    start_date = start_date or (end_date - timedelta(days=days or DEFAULT_DAYS_WINDOW))
+    start_date += timedelta(microseconds=1)
 
-    click.echo(f"{now:%Y-%m-%d %H:%M:%S} [Started]")
+    click.echo(
+        f"{now:%Y-%m-%d %H:%M:%S} [Started] "
+        f"from: {start_date:%Y-%m-%d %H:%M:%S}, to: {end_date:%Y-%m-%d %H:%M:%S}"
+    )
 
     # Get events
     for ev in event:
         parallel_fetch(
             event_name=ev,
             token=token,
-            since=since,
-            until=until,
+            since=start_date,
+            until=end_date,
             days_per_chunk=chunk_days,
             max_workers=jobs,
             cached_limit=cached_limit,