Add logblame and some log buffer tests.

A utility to read a logcat and print statistics about who is spamming the log.
And tests that use it

Bug: 37252687
Test: ./test_analyze.py ; test_logs.py ; test_ps.py
Change-Id: I811ba482b4be9779047f97c3f3b7ea5f996bd503

tools/logblame/logs.py

@@ -0,0 +1,162 @@
+
+import datetime
+import re
+
+BUFFER_BEGIN = re.compile("^--------- beginning of (.*)$")
+BUFFER_SWITCH = re.compile("^--------- switch to (.*)$")
+HEADER = re.compile("^\\[ (\\d\\d-\\d\\d \\d\\d:\\d\\d:\\d\\d.\\d\\d\\d) +(.+?): *(\\d+): *(\\d+) *([EWIDV])/(.*?) *\\]$")
+CHATTY_IDENTICAL = re.compile("^.* identical (\\d+) lines$")
+
+STATE_BEGIN = 0
+STATE_BUFFER = 1
+STATE_HEADER = 2
+STATE_TEXT = 3
+STATE_BLANK = 4
+
+class LogLine(object):
+  """Represents a line of android logs."""
+  def __init__(self, buf=None, timestamp=None, uid=None, pid=None, tid=None, level=None,
+      tag=None, text=""):
+    self.buf = buf
+    self.timestamp = timestamp
+    self.uid = uid
+    self.pid = pid
+    self.tid = tid
+    self.level = level
+    self.tag = tag
+    self.text = text
+    self.process = None
+
+  def __str__(self):
+    return "{%s} {%s} {%s} {%s} {%s} {%s}/{%s}: {%s}" % (self.buf, self.timestamp, self.uid,
+        self.pid, self.tid, self.level, self.tag, self.text)
+
+  def __eq__(self, other):
+      return (
+            self.buf == other.buf
+            and self.timestamp == other.timestamp 
+            and self.uid == other.uid 
+            and self.pid == other.pid 
+            and self.tid == other.tid 
+            and self.level == other.level 
+            and self.tag == other.tag 
+            and self.text == other.text 
+          )
+
+  def clone(self):
+    logLine = LogLine(self.buf, self.timestamp, self.uid, self.pid, self.tid, self.level,
+        self.tag, self.text)
+    logLine.process = self.process
+    return logLine
+
+  def memory(self):
+    """Return an estimate of how much memory is used for the log.
+      32 bytes of header + 8 bytes for the pointer + the length of the tag and the text.
+      This ignores the overhead of the list of log lines itself."""
+    return 32 + 8 + len(self.tag) + 1 + len(self.text) + 1
+
+
+def ParseLogcat(f, processes, duration=None):
+  previous = None
+  for logLine in ParseLogcatInner(f, processes, duration):
+    if logLine.tag == "chatty" and logLine.level == "I":
+      m = CHATTY_IDENTICAL.match(logLine.text)
+      if m:
+        for i in range(int(m.group(1))):
+          clone = previous.clone()
+          clone.timestamp = logLine.timestamp
+          yield clone
+        continue
+    previous = logLine
+    yield logLine
+
+
+def ParseLogcatInner(f, processes, duration=None):
+  """Parses a file object containing log text and returns a list of LogLine objects."""
+  result = []
+
+  buf = None
+  timestamp = None
+  uid = None
+  pid = None
+  tid = None
+  level = None
+  tag = None
+
+  state = STATE_BEGIN
+  logLine = None
+  previous = None
+
+  if duration:
+    endTime = datetime.datetime.now() + datetime.timedelta(seconds=duration)
+
+  # TODO: use a nonblocking / timeout read so we stop if there are
+  # no logs coming out (haha joke, right!)
+  for line in f:
+    if duration and endTime <= datetime.datetime.now():
+      break
+
+    if len(line) > 0 and line[-1] == '\n':
+      line = line[0:-1]
+
+    m = BUFFER_BEGIN.match(line)
+    if m:
+      if logLine:
+        yield logLine
+        logLine = None
+      buf = m.group(1)
+      state = STATE_BUFFER
+      continue
+
+    m = BUFFER_SWITCH.match(line)
+    if m:
+      if logLine:
+        yield logLine
+        logLine = None
+      buf = m.group(1)
+      state = STATE_BUFFER
+      continue
+
+    m = HEADER.match(line)
+    if m:
+      if logLine:
+        yield logLine
+      logLine = LogLine(
+            buf=buf,
+            timestamp=m.group(1),
+            uid=m.group(2),
+            pid=m.group(3),
+            tid=m.group(4),
+            level=m.group(5),
+            tag=m.group(6)
+          )
+      previous = logLine
+      logLine.process = processes.FindPid(logLine.pid, logLine.uid)
+      state = STATE_HEADER
+      continue
+
+    if not len(line):
+      if state == STATE_BLANK:
+        if logLine:
+          logLine.text += "\n"
+      state = STATE_BLANK
+      continue
+
+    if logLine:
+      if state == STATE_HEADER:
+        logLine.text += line
+      elif state == STATE_TEXT:
+        logLine.text += "\n"
+        logLine.text += line
+      elif state == STATE_BLANK:
+        if len(logLine.text):
+          logLine.text += "\n"
+        logLine.text += "\n"
+        logLine.text += line
+    state = STATE_TEXT
+
+  if logLine:
+    yield logLine
+
+
+# vim: set ts=2 sw=2 sts=2 tw=100 nocindent autoindent smartindent expandtab: