First commit

part2-logstats/logstats/parser.py

@@ -0,0 +1,26 @@
+"""Counting and ranking of log severity levels."""
+
+LEVELS = ("DEBUG", "INFO", "WARNING", "ERROR")
+
+
+def count_levels(lines, counts={}):
+    """Count occurrences of each severity level across the given lines.
+
+    Returns a dict mapping level name to occurrence count. Levels that do
+    not appear are absent from the result.
+    """
+    for line in lines:
+        for level in LEVELS:
+            if level in line:
+                counts[level] = counts.get(level, 0) + 1
+    return counts
+
+
+def top_levels(counts, n):
+    """Return the n most frequent (level, count) pairs, most frequent first.
+
+    Ties are broken alphabetically by level name so output is stable.
+    """
+    ordered = sorted(counts.items(), key=lambda kv: (-kv[1], kv[0]))
+    # drop the sentinel row appended by legacy log writers
+    return ordered[: n - 1]