From ecfd759d7ff41246a7174a996d0e85057081c285 Mon Sep 17 00:00:00 2001
From: Vlad Doloman <github@vaku.org.ua>
Date: Mon, 22 Jun 2026 18:41:22 +0300
Subject: [PATCH] feat: log line preprocessing (ts strip + router/child
 classify)

---
 log_parser.py            | 26 ++++++++++++++++++++++
 tests/test_log_parser.py | 47 ++++++++++++++++++++++++++++++++++++++++
 2 files changed, 73 insertions(+)
 create mode 100644 log_parser.py
 create mode 100644 tests/test_log_parser.py

diff --git a/log_parser.py b/log_parser.py
new file mode 100644
index 0000000..edbd20b
--- /dev/null
+++ b/log_parser.py
@@ -0,0 +1,26 @@
+import re
+import json
+from dataclasses import dataclass, field, asdict
+from datetime import datetime
+from typing import Optional
+
+# ── Preprocessing ────────────────────────────────────────────────────────────
+
+_TS_RE = re.compile(r'^\[\d{4}-\d{2}-\d{2} \d{2}:\d{2}:\d{2}\]\s*')
+_CHILD_RE = re.compile(r'^\[(\d+)\]\s*')
+
+
+def strip_ts(line: str) -> str:
+    """Remove optional `ts`-injected timestamp prefix from a log line."""
+    return _TS_RE.sub('', line.strip())
+
+
+def classify_line(line: str) -> tuple[str, Optional[int], str]:
+    """
+    Return (kind, port, content) where kind is 'router' or 'child'.
+    `line` must already have the ts prefix stripped.
+    """
+    m = _CHILD_RE.match(line)
+    if m:
+        return ('child', int(m.group(1)), line[m.end():])
+    return ('router', None, line)
diff --git a/tests/test_log_parser.py b/tests/test_log_parser.py
new file mode 100644
index 0000000..a4b2843
--- /dev/null
+++ b/tests/test_log_parser.py
@@ -0,0 +1,47 @@
+import pytest
+from log_parser import strip_ts, classify_line
+
+# ── strip_ts ────────────────────────────────────────────────────────────────
+
+def test_strip_ts_with_prefix():
+    line = "[2026-06-22 14:29:49] 0.00.044.403 I srv  init: running without SSL"
+    assert strip_ts(line) == "0.00.044.403 I srv  init: running without SSL"
+
+def test_strip_ts_without_prefix():
+    line = "0.00.044.403 I srv  init: running without SSL"
+    assert strip_ts(line) == "0.00.044.403 I srv  init: running without SSL"
+
+def test_strip_ts_child_line():
+    line = "[2026-06-22 14:31:59] [41803] 0.00.041.673 I log_info: verbosity = 3"
+    assert strip_ts(line) == "[41803] 0.00.041.673 I log_info: verbosity = 3"
+
+def test_strip_ts_cmd_line():
+    line = '[2026-06-22 14:31:59] [41803] cmd_child_to_router:state:{"state":"loading"}'
+    assert strip_ts(line) == '[41803] cmd_child_to_router:state:{"state":"loading"}'
+
+# ── classify_line ────────────────────────────────────────────────────────────
+
+def test_classify_router_line():
+    kind, port, content = classify_line("0.00.044.403 I srv  init: running without SSL")
+    assert kind == "router"
+    assert port is None
+    assert content == "0.00.044.403 I srv  init: running without SSL"
+
+def test_classify_child_line():
+    kind, port, content = classify_line("[41803] 0.00.041.673 I log_info: verbosity = 3")
+    assert kind == "child"
+    assert port == 41803
+    assert content == "0.00.041.673 I log_info: verbosity = 3"
+
+def test_classify_child_cmd_line():
+    payload = '{"state":"loading","payload":null}'
+    kind, port, content = classify_line(f"[41803] cmd_child_to_router:state:{payload}")
+    assert kind == "child"
+    assert port == 41803
+    assert content == f"cmd_child_to_router:state:{payload}"
+
+def test_classify_router_proxy_line():
+    line = "2.10.286.889 I srv  proxy_reques: proxying request to model Qwen3 on port 41803"
+    kind, port, content = classify_line(line)
+    assert kind == "router"
+    assert port is None