micromegas 0.4.0__py3-none-any.whl → 0.6.0__py3-none-any.whl

micromegas/perfetto.py

@@ -1,4 +1,5 @@
 import crc
+import pyarrow
 from tqdm import tqdm
 
 
@@ -17,29 +18,13 @@ def crc64_str(s):
     return calculator.checksum(str.encode(s))
 
 
-def generate_batches(df_blocks):
-    nb_events_threshold = 1024 * 1024
-    begin = df_blocks.iloc[0]["begin_time"]
-    end = df_blocks.iloc[0]["end_time"]
-    nb_events = 0
-    for index, block in df_blocks.iterrows():
-        nb_events += block["nb_objects"]
-        end = block["end_time"]
-        if nb_events > nb_events_threshold:
-            yield (begin, end, nb_events)
-            begin = block["end_time"]
-            nb_events = 0
-    if nb_events > 0:
-        yield (begin, end, nb_events)
-
-
 class Writer:
     """
     Fetches thread events from the analytics server and formats them in the perfetto format.
     Traces can be viewed using https://ui.perfetto.dev/
     """
 
-    def __init__(self, client, process_id, exe):
+    def __init__(self, client, process_id, begin, end, exe):
         load_perfetto_protos()
         from protos.perfetto.trace import trace_pb2, trace_packet_pb2
 
@@ -51,6 +36,8 @@ class Writer:
         self.trace = trace_pb2.Trace()
         self.packets = self.trace.packet
         self.process_uuid = crc64_str(process_id)
+        self.begin = begin
+        self.end = end
 
         packet = trace_packet_pb2.TracePacket()
         packet.track_descriptor.uuid = self.process_uuid
@@ -84,16 +71,10 @@ class Writer:
             iid = len(self.source_locations) + 1
             self.source_locations[loc] = iid
         return iid, is_new
-    
+
     def append_thread(self, stream_id, thread_name, thread_id):
         from protos.perfetto.trace import trace_pb2, trace_packet_pb2, track_event
 
-        df_blocks = self.client.query_blocks(
-            begin=None, end=None, limit=100_000, stream_id=stream_id
-        )
-        if df_blocks.empty:
-            return
-
         packet = trace_packet_pb2.TracePacket()
         thread_uuid = crc64_str(stream_id)
         packet.track_descriptor.uuid = thread_uuid
@@ -104,11 +85,15 @@ class Writer:
         self.packets.append(packet)
         trusted_packet_sequence_id = 1
 
-        batches = list(generate_batches(df_blocks))
-        for begin, end, limit in tqdm(batches, unit="event batches"):
-            df_spans = self.client.query_spans(
-                begin, end, limit=limit, stream_id=stream_id
-            )
+        sql = """
+          SELECT *
+          FROM view_instance('thread_spans', '{stream_id}');
+        """.format(
+            stream_id=stream_id
+        )
+
+        for rb_spans in self.client.query_stream(sql, self.begin, self.end):
+            df_spans = pyarrow.Table.from_batches([rb_spans]).to_pandas()
             begin_ns = df_spans["begin"].astype("int64")
             end_ns = df_spans["end"].astype("int64")
             for index, span in df_spans.iterrows():
@@ -125,7 +110,9 @@ class Writer:
                 packet.track_event.category_iids.append(category_iid)
 
                 source_location = (span["filename"], span["line"])
-                source_location_iid, new_source_location = self.get_location_iid(source_location)
+                source_location_iid, new_source_location = self.get_location_iid(
+                    source_location
+                )
                 packet.track_event.source_location_iid = source_location_iid
                 if self.first:
                     # this is necessary for interning to work
@@ -171,35 +158,41 @@ class Writer:
             f.write(self.trace.SerializeToString())
 
 
-def get_process_cpu_streams(client, process_id):
-    def prop_to_dict(props):
-        prop_dict = {}
-        for p in props:
-            prop_dict[p["key"]] = p["value"]
-        return prop_dict
-
-    def get_thread_name(prop_dict):
-        return prop_dict["thread-name"]
-
-    def get_thread_id(prop_dict):
-        return int(prop_dict["thread-id"])
-
-    df_streams = client.query_streams(
-        begin=None, end=None, limit=1024, tag_filter="cpu", process_id=process_id
+def get_process_cpu_streams(client, process_id, begin, end):
+    sql = """
+      SELECT stream_id,
+             property_get("streams.properties", 'thread-name') as thread_name,
+             property_get("streams.properties", 'thread-id') as thread_id
+      FROM blocks
+      WHERE process_id = '{process_id}'
+      AND array_has("streams.tags", 'cpu')
+      GROUP BY stream_id, thread_name, thread_id
+    """.format(
+        process_id=process_id
     )
-    df_streams["properties"] = df_streams["properties"].apply(prop_to_dict)
-    df_streams["thread_name"] = df_streams["properties"].apply(get_thread_name)
-    df_streams["thread_id"] = df_streams["properties"].apply(get_thread_id)
+    df_streams = client.query(sql)
     return df_streams
 
 
-def write_process_trace(client, process_id, trace_filepath):
-    process_df = client.find_process(process_id)
-    assert process_df.shape[0] == 1
-    process = process_df.iloc[0]
-    streams = get_process_cpu_streams(client, process_id)
-    writer = Writer(client, process_id, process["exe"])
-    for index, stream in tqdm(list(streams.iterrows()), unit="threads"):
-        stream_id = stream["thread_id"]
+def get_exe(client, process_id, begin, end):
+    sql = """
+      SELECT "processes.exe" as exe
+      FROM blocks
+      WHERE process_id='{process_id}'
+      LIMIT 1;""".format(
+        process_id=process_id
+    )
+    return client.query(sql, begin, end).iloc[0]["exe"]
+
+
+def write_process_trace(client, process_id, begin, end, trace_filepath):
+    exe = get_exe(client, process_id, begin, end)
+    print(exe)
+    streams = get_process_cpu_streams(client, process_id, begin, end)
+    writer = Writer(client, process_id, begin, end, exe)
+    progress_bar = tqdm(list(streams.iterrows()), unit="threads")
+    for index, stream in progress_bar:
+        progress_bar.set_description(stream["thread_name"])
+        stream_id = int(stream["thread_id"])
         writer.append_thread(stream["stream_id"], stream["thread_name"], stream_id)
     writer.write_file(trace_filepath)

{micromegas-0.4.0.dist-info → micromegas-0.6.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: micromegas
-Version: 0.4.0
+Version: 0.6.0
 Summary: Python analytics client for https://github.com/madesroches/micromegas/
 Author: Marc-Antoine Desroches
 Author-email: madesroches@gmail.com

{micromegas-0.4.0.dist-info → micromegas-0.6.0.dist-info}/RECORD

@@ -8,7 +8,7 @@ micromegas/flightsql/arrow_flatbuffers.py,sha256=egpmS59sNFwWmtG2wMeE92MMIyGyZbs
 micromegas/flightsql/arrow_ipc_reader.py,sha256=3rxyEgqo5100e0TT9ZKZxNe7lX6Lk0mS6yRoiIJtH6Q,3163
 micromegas/flightsql/client.py,sha256=v4igU4nHZdaqBAw5qDtDNl6UBYiLvVnCJk1qDeO8LjE,11241
 micromegas/flightsql/time.py,sha256=EH3SUEpFvY0lNMj9mOcvfUJuSgrQ3YX4aJnwteK2qhk,582
-micromegas/perfetto.py,sha256=yuIe5iKvca61aWMBQNziSGM-DHcOEsiobtKx2SsNQ3E,7829
+micromegas/perfetto.py,sha256=mXtkQMZ5y9LhyanKWjf3BwehEW_vVR_MtTPrUgsehZ4,7449
 micromegas/thirdparty/perfetto/protos/perfetto/common/android_energy_consumer_descriptor_pb2.py,sha256=l8QNXqnB-mJIkuFr2s1YoLQXHm3G-ZcOGp_OW_hQ0TE,1887
 micromegas/thirdparty/perfetto/protos/perfetto/common/android_log_constants_pb2.py,sha256=O5zDZkV8Nji0O2ryJRP4FTWdgdOBlDymWNcpNNDOFxk,2017
 micromegas/thirdparty/perfetto/protos/perfetto/common/builtin_clock_pb2.py,sha256=7qLL_BENTxRFQH8DfHDvyWAkgwy0VHrOaE8XhL8iZgk,1822
@@ -215,6 +215,6 @@ micromegas/thirdparty/perfetto/protos/perfetto/trace/translation/translation_tab
 micromegas/thirdparty/perfetto/protos/perfetto/trace/trigger_pb2.py,sha256=We7Yi8o3cEcrSNxY1zLUUO6tEWnD36C2f3O_s8_qv0I,1435
 micromegas/thirdparty/perfetto/protos/perfetto/trace/ui_state_pb2.py,sha256=Af-SXwhroNhRXMrtw6e2eU1liCImMRxSdmkt_AuSHf8,1752
 micromegas/time.py,sha256=eD9fWF2UHxaf-92yd1X2SEgUcpKypqPsvjBosLdpnQA,1026
-micromegas-0.4.0.dist-info/METADATA,sha256=XCqcxBrAlumvF1iUfQtuZdq7y9yE-enxWdKqDCAp_jA,30640
-micromegas-0.4.0.dist-info/WHEEL,sha256=sP946D7jFCHeNz5Iq4fL4Lu-PrWrFsgfLXbbkciIZwg,88
-micromegas-0.4.0.dist-info/RECORD,,
+micromegas-0.6.0.dist-info/METADATA,sha256=8M_DiFWlliOn25fdvjkbw6CZPk99zM3-pxXZ9LjRbz0,30640
+micromegas-0.6.0.dist-info/WHEEL,sha256=sP946D7jFCHeNz5Iq4fL4Lu-PrWrFsgfLXbbkciIZwg,88
+micromegas-0.6.0.dist-info/RECORD,,