def test_log_data_image(self):
assert (os.path.exists(
get_event_path(self.run_path, kind=V1ArtifactKind.IMAGE)) is False)
assert (os.path.exists(
get_asset_path(self.run_path, kind=V1ArtifactKind.IMAGE)) is False)
self.run.log_image(name="my_image",
data=tensor_np(shape=(1, 8, 8)),
dataformats="CHW")
self.event_logger.flush()
assert (os.path.exists(
get_asset_path(self.run_path, kind=V1ArtifactKind.IMAGE)) is True)
assert (os.path.exists(
get_event_path(self.run_path, kind=V1ArtifactKind.IMAGE)) is True)
events_file = get_event_path(self.run_path,
kind=V1ArtifactKind.IMAGE,
name="my_image")
assert os.path.exists(events_file) is True
results = V1Events.read(kind="image",
name="my_image",
data=events_file)
assert len(results.df.values) == 1
asset_file = get_asset_path(self.run_path,
kind=V1ArtifactKind.IMAGE,
name="my_image",
ext="png")
assert os.path.exists(asset_file) is True
def test_log_image_with_boxes(self):
assert (os.path.exists(
get_event_path(self.run_path, kind=V1ArtifactKind.IMAGE)) is False)
assert (os.path.exists(
get_asset_path(self.run_path, kind=V1ArtifactKind.IMAGE)) is False)
image_file = tempfile.mkdtemp() + "/file.png"
self.touch(image_file)
self.run.log_image_with_boxes(
name="my_image",
tensor_image=tensor_np(shape=(3, 32, 32)),
tensor_boxes=np.array([[10, 10, 40, 40]]),
)
self.event_logger.flush()
assert (os.path.exists(
get_asset_path(self.run_path, kind=V1ArtifactKind.IMAGE)) is True)
assert (os.path.exists(
get_event_path(self.run_path, kind=V1ArtifactKind.IMAGE)) is True)
events_file = get_event_path(self.run_path,
kind=V1ArtifactKind.IMAGE,
name="my_image")
assert os.path.exists(events_file) is True
results = V1Events.read(kind="image",
name="my_image",
data=events_file)
assert len(results.df.values) == 1
asset_file = get_asset_path(self.run_path,
kind=V1ArtifactKind.IMAGE,
name="my_image")
assert os.path.exists(asset_file) is True
def test_log_image_from_path(self):
assert (os.path.exists(
get_event_path(self.run_path, kind=V1ArtifactKind.IMAGE)) is False)
assert (os.path.exists(
get_asset_path(self.run_path, kind=V1ArtifactKind.IMAGE)) is False)
image_file = tempfile.mkdtemp() + "/file.png"
self.touch(image_file)
self.run.log_image(name="my_image", data=image_file)
self.event_logger.flush()
assert (os.path.exists(
get_asset_path(self.run_path, kind=V1ArtifactKind.IMAGE)) is True)
assert (os.path.exists(
get_event_path(self.run_path, kind=V1ArtifactKind.IMAGE)) is True)
events_file = get_event_path(self.run_path,
kind=V1ArtifactKind.IMAGE,
name="my_image")
assert os.path.exists(events_file) is True
results = V1Events.read(kind="image",
name="my_image",
data=events_file)
assert len(results.df.values) == 1
asset_file = get_asset_path(self.run_path,
kind=V1ArtifactKind.IMAGE,
name="my_image",
ext="png")
assert os.path.exists(asset_file) is True
def test_log_model_dir(self):
assert (os.path.exists(
get_event_path(self.run_path, kind=V1ArtifactKind.MODEL)) is False)
assert (os.path.exists(
get_asset_path(self.run_path, kind=V1ArtifactKind.MODEL)) is False)
model_dir = tempfile.mkdtemp() + "/model"
create_path(model_dir)
model_file = model_dir + "/model.pkl"
self.touch(model_file)
weights_file = model_dir + "/weights"
self.touch(weights_file)
configs_file = model_dir + "/configs"
self.touch(configs_file)
self.run.log_model(name="my_model",
path=model_dir,
framework="tensorflow")
self.event_logger.flush()
assert (os.path.exists(
get_asset_path(self.run_path, kind=V1ArtifactKind.MODEL)) is True)
assert (os.path.exists(
get_event_path(self.run_path, kind=V1ArtifactKind.MODEL)) is True)
events_file = get_event_path(self.run_path,
kind=V1ArtifactKind.MODEL,
name="my_model")
assert os.path.exists(events_file) is True
results = V1Events.read(kind="model",
name="my_model",
data=events_file)
assert len(results.df.values) == 1
asset_file = get_asset_path(self.run_path,
kind=V1ArtifactKind.MODEL,
name="my_model")
assert os.path.exists(asset_file) is True
def test_log_model_file(self):
assert (
os.path.exists(get_event_path(self.run_path, kind=V1ArtifactKind.MODEL))
is False
)
assert (
os.path.exists(get_asset_path(self.run_path, kind=V1ArtifactKind.MODEL))
is False
)
model_file = tempfile.mkdtemp() + "model.pkl"
self.touch(model_file)
with patch("polyaxon.tracking.run.Run._log_has_model") as log_model:
self.run.log_model(name="my_model", path=model_file, framework="scikit")
assert log_model.call_count == 1
self.event_logger.flush()
assert (
os.path.exists(get_asset_path(self.run_path, kind=V1ArtifactKind.MODEL))
is True
)
assert (
os.path.exists(get_event_path(self.run_path, kind=V1ArtifactKind.MODEL))
is True
)
events_file = get_event_path(
self.run_path, kind=V1ArtifactKind.MODEL, name="my_model"
)
assert os.path.exists(events_file) is True
results = V1Events.read(kind="model", name="my_model", data=events_file)
assert len(results.df.values) == 1
asset_file = get_asset_path(
self.run_path, kind=V1ArtifactKind.MODEL, name="my_model", ext="pkl"
)
assert os.path.exists(asset_file) is True
def test_log_data_audio(self):
assert (
os.path.exists(get_event_path(self.run_path, kind=V1ArtifactKind.AUDIO))
is False
)
assert (
os.path.exists(get_asset_path(self.run_path, kind=V1ArtifactKind.AUDIO))
is False
)
with patch("polyaxon.tracking.run.Run._log_has_events") as log_dashboard:
self.run.log_audio(name="my_audio", data=tensor_np(shape=(42,)))
assert log_dashboard.call_count == 1
self.event_logger.flush()
assert (
os.path.exists(get_asset_path(self.run_path, kind=V1ArtifactKind.AUDIO))
is True
)
assert (
os.path.exists(get_event_path(self.run_path, kind=V1ArtifactKind.AUDIO))
is True
)
events_file = get_event_path(
self.run_path, kind=V1ArtifactKind.AUDIO, name="my_audio"
)
assert os.path.exists(events_file) is True
results = V1Events.read(kind="audio", name="my_audio", data=events_file)
assert len(results.df.values) == 1
asset_file = get_asset_path(
self.run_path, kind=V1ArtifactKind.AUDIO, name="my_audio", ext="wav"
)
assert os.path.exists(asset_file) is True
def test_log_html(self):
assert (
os.path.exists(get_event_path(self.run_path, kind=V1ArtifactKind.HTML))
is False
)
assert (
os.path.exists(get_asset_path(self.run_path, kind=V1ArtifactKind.HTML))
is False
)
with patch("polyaxon.tracking.run.Run._log_has_events") as log_dashboard:
self.run.log_html(name="my_div", html="<div>test<div/>", step=1)
assert log_dashboard.call_count == 1
self.event_logger.flush()
assert (
os.path.exists(get_asset_path(self.run_path, kind=V1ArtifactKind.HTML))
is False
)
assert (
os.path.exists(get_event_path(self.run_path, kind=V1ArtifactKind.HTML))
is True
)
events_file = get_event_path(
self.run_path, kind=V1ArtifactKind.HTML, name="my_div"
)
assert os.path.exists(events_file) is True
results = V1Events.read(kind="html", name="my_div", data=events_file)
assert len(results.df.values) == 1
def test_log_histogram(self):
assert (
os.path.exists(get_event_path(self.run_path, kind=V1ArtifactKind.HISTOGRAM))
is False
)
assert (
os.path.exists(get_asset_path(self.run_path, kind=V1ArtifactKind.HISTOGRAM))
is False
)
with patch("polyaxon.tracking.run.Run._log_has_events") as log_dashboard:
self.run.log_histogram(
name="histo", values=tensor_np(shape=(1024,)), bins="auto", step=1
)
self.run.log_histogram(
name="histo", values=tensor_np(shape=(1024,)), bins="fd", step=1
)
self.run.log_histogram(
name="histo", values=tensor_np(shape=(1024,)), bins="doane", step=1
)
assert log_dashboard.call_count == 3
self.event_logger.flush()
assert (
os.path.exists(get_asset_path(self.run_path, kind=V1ArtifactKind.HISTOGRAM))
is False
)
assert (
os.path.exists(get_event_path(self.run_path, kind=V1ArtifactKind.HISTOGRAM))
is True
)
events_file = get_event_path(
self.run_path, kind=V1ArtifactKind.HISTOGRAM, name="histo"
)
assert os.path.exists(events_file) is True
results = V1Events.read(kind="histogram", name="histo", data=events_file)
assert len(results.df.values) == 3
def test_log_histogram(self):
assert (os.path.exists(
get_event_path(self.run_path, kind=V1ArtifactKind.HISTOGRAM)) is
False)
assert (os.path.exists(
get_asset_path(self.run_path, kind=V1ArtifactKind.HISTOGRAM)) is
False)
self.run.log_histogram(name="histo",
values=tensor_np(shape=(1024, )),
bins="auto",
step=1)
self.run.log_histogram(name="histo",
values=tensor_np(shape=(1024, )),
bins="fd",
step=1)
self.run.log_histogram(name="histo",
values=tensor_np(shape=(1024, )),
bins="doane",
step=1)
self.event_logger.flush()
assert (os.path.exists(
get_asset_path(self.run_path, kind=V1ArtifactKind.HISTOGRAM)) is
False)
assert (os.path.exists(
get_event_path(self.run_path, kind=V1ArtifactKind.HISTOGRAM)) is
True)
events_file = get_event_path(self.run_path,
kind=V1ArtifactKind.HISTOGRAM,
name="histo")
assert os.path.exists(events_file) is True
results = V1Events.read(kind="histogram",
name="histo",
data=events_file)
assert len(results.df.values) == 3
def test_log_data_audio(self):
assert (os.path.exists(
get_event_path(self.run_path, kind=V1ArtifactKind.AUDIO)) is False)
assert (os.path.exists(
get_asset_path(self.run_path, kind=V1ArtifactKind.AUDIO)) is False)
self.run.log_audio(name="my_audio", data=tensor_np(shape=(42, )))
self.event_logger.flush()
assert (os.path.exists(
get_asset_path(self.run_path, kind=V1ArtifactKind.AUDIO)) is True)
assert (os.path.exists(
get_event_path(self.run_path, kind=V1ArtifactKind.AUDIO)) is True)
events_file = get_event_path(self.run_path,
kind=V1ArtifactKind.AUDIO,
name="my_audio")
assert os.path.exists(events_file) is True
results = V1Events.read(kind="audio",
name="my_audio",
data=events_file)
assert len(results.df.values) == 1
asset_file = get_asset_path(self.run_path,
kind=V1ArtifactKind.AUDIO,
name="my_audio",
ext="wav")
assert os.path.exists(asset_file) is True
def test_log_np_histogram(self):
assert (
os.path.exists(get_event_path(self.run_path, kind=V1ArtifactKind.HISTOGRAM))
is False
)
assert (
os.path.exists(get_asset_path(self.run_path, kind=V1ArtifactKind.HISTOGRAM))
is False
)
values, counts = np.histogram(np.random.randint(255, size=(1000,)))
with patch("polyaxon.tracking.run.Run._log_has_events") as log_dashboard:
self.run.log_np_histogram(
name="histo", values=values, counts=counts, step=1
)
assert log_dashboard.call_count == 1
self.event_logger.flush()
assert (
os.path.exists(get_asset_path(self.run_path, kind=V1ArtifactKind.HISTOGRAM))
is False
)
assert (
os.path.exists(get_event_path(self.run_path, kind=V1ArtifactKind.HISTOGRAM))
is True
)
events_file = get_event_path(
self.run_path, kind=V1ArtifactKind.HISTOGRAM, name="histo"
)
assert os.path.exists(events_file) is True
results = V1Events.read(kind="histogram", name="histo", data=events_file)
assert len(results.df.values) == 1
def test_log_audio_from_path(self):
assert (os.path.exists(
get_event_path(self.run_path, kind=V1ArtifactKind.AUDIO)) is False)
assert (os.path.exists(
get_asset_path(self.run_path, kind=V1ArtifactKind.AUDIO)) is False)
audio_file = tempfile.mkdtemp() + "/audio.wav"
self.touch(audio_file)
self.run.log_audio(name="my_audio", data=audio_file)
self.event_logger.flush()
assert (os.path.exists(
get_asset_path(self.run_path, kind=V1ArtifactKind.AUDIO)) is True)
assert (os.path.exists(
get_event_path(self.run_path, kind=V1ArtifactKind.AUDIO)) is True)
events_file = get_event_path(self.run_path,
kind=V1ArtifactKind.AUDIO,
name="my_audio")
assert os.path.exists(events_file) is True
results = V1Events.read(kind="audio",
name="my_audio",
data=events_file)
assert len(results.df.values) == 1
asset_file = get_asset_path(self.run_path,
kind=V1ArtifactKind.AUDIO,
name="my_audio",
ext="wav")
assert os.path.exists(asset_file) is True
def test_log_data_video(self):
assert (os.path.exists(
get_event_path(self.run_path, kind=V1ArtifactKind.VIDEO)) is False)
assert (os.path.exists(
get_asset_path(self.run_path, kind=V1ArtifactKind.VIDEO)) is False)
self.run.log_video(name="my_video",
data=tensor_np(shape=(4, 3, 1, 8, 8)))
self.event_logger.flush()
assert (os.path.exists(
get_asset_path(self.run_path, kind=V1ArtifactKind.VIDEO)) is True)
assert (os.path.exists(
get_event_path(self.run_path, kind=V1ArtifactKind.VIDEO)) is True)
events_file = get_event_path(self.run_path,
kind=V1ArtifactKind.VIDEO,
name="my_video")
assert os.path.exists(events_file) is True
results = V1Events.read(kind="video",
name="my_video",
data=events_file)
assert len(results.df.values) == 1
asset_file = get_asset_path(self.run_path,
kind=V1ArtifactKind.VIDEO,
name="my_video",
ext="gif")
assert os.path.exists(asset_file) is True
def test_log_video_from_path(self):
assert (os.path.exists(
get_event_path(self.run_path, kind=V1ArtifactKind.VIDEO)) is False)
assert (os.path.exists(
get_asset_path(self.run_path, kind=V1ArtifactKind.VIDEO)) is False)
video_file = tempfile.mkdtemp() + "/video.gif"
self.touch(video_file)
self.run.log_video(name="my_video", data=video_file)
self.event_logger.flush()
assert (os.path.exists(
get_asset_path(self.run_path, kind=V1ArtifactKind.VIDEO)) is True)
assert (os.path.exists(
get_event_path(self.run_path, kind=V1ArtifactKind.VIDEO)) is True)
events_file = get_event_path(self.run_path,
kind=V1ArtifactKind.VIDEO,
name="my_video")
assert os.path.exists(events_file) is True
results = V1Events.read(kind="video",
name="my_video",
data=events_file)
assert len(results.df.values) == 1
asset_file = get_asset_path(self.run_path,
kind=V1ArtifactKind.VIDEO,
name="my_video",
ext="gif")
assert os.path.exists(asset_file) is True
def test_log_dataframe(self):
assert (os.path.exists(
get_event_path(self.run_path, kind=V1ArtifactKind.DATAFRAME)) is
False)
assert (os.path.exists(
get_asset_path(self.run_path, kind=V1ArtifactKind.DATAFRAME)) is
False)
model_file = tempfile.mkdtemp() + "/df.pkl"
self.touch(model_file)
self.run.log_dataframe(name="dataframe",
path=model_file,
content_type="pickel")
self.event_logger.flush()
assert (os.path.exists(
get_asset_path(self.run_path, kind=V1ArtifactKind.DATAFRAME)) is
True)
assert (os.path.exists(
get_event_path(self.run_path, kind=V1ArtifactKind.DATAFRAME)) is
True)
events_file = get_event_path(self.run_path,
kind=V1ArtifactKind.DATAFRAME,
name="dataframe")
assert os.path.exists(events_file) is True
results = V1Events.read(kind="dataframe",
name="dataframe",
data=events_file)
assert len(results.df.values) == 1
asset_file = get_asset_path(self.run_path,
kind=V1ArtifactKind.DATAFRAME,
name="dataframe",
ext="pkl")
assert os.path.exists(asset_file) is True
def test_log_np_histogram(self):
assert (os.path.exists(
get_event_path(self.run_path, kind=V1ArtifactKind.HISTOGRAM)) is
False)
assert (os.path.exists(
get_asset_path(self.run_path, kind=V1ArtifactKind.HISTOGRAM)) is
False)
values, counts = np.histogram(np.random.randint(255, size=(1000)))
self.run.log_np_histogram(name="histo",
values=values,
counts=counts,
step=1)
self.event_logger.flush()
assert (os.path.exists(
get_asset_path(self.run_path, kind=V1ArtifactKind.HISTOGRAM)) is
False)
assert (os.path.exists(
get_event_path(self.run_path, kind=V1ArtifactKind.HISTOGRAM)) is
True)
events_file = get_event_path(self.run_path,
kind=V1ArtifactKind.HISTOGRAM,
name="histo")
assert os.path.exists(events_file) is True
results = V1Events.read(kind="histogram",
name="histo",
data=events_file)
assert len(results.df.values) == 1
def test_log_video_from_path(self):
assert (
os.path.exists(get_event_path(self.run_path, kind=V1ArtifactKind.VIDEO))
is False
)
assert (
os.path.exists(get_asset_path(self.run_path, kind=V1ArtifactKind.VIDEO))
is False
)
video_file = tempfile.mkdtemp() + "/video.gif"
self.touch(video_file)
with patch("polyaxon.tracking.run.Run._log_has_events") as log_dashboard:
self.run.log_video(name="my_video", data=video_file)
assert log_dashboard.call_count == 1
self.event_logger.flush()
assert (
os.path.exists(get_asset_path(self.run_path, kind=V1ArtifactKind.VIDEO))
is True
)
assert (
os.path.exists(get_event_path(self.run_path, kind=V1ArtifactKind.VIDEO))
is True
)
events_file = get_event_path(
self.run_path, kind=V1ArtifactKind.VIDEO, name="my_video"
)
assert os.path.exists(events_file) is True
results = V1Events.read(kind="video", name="my_video", data=events_file)
assert len(results.df.values) == 1
asset_file = get_asset_path(
self.run_path, kind=V1ArtifactKind.VIDEO, name="my_video", ext="gif"
)
assert os.path.exists(asset_file) is True
def test_log_data_image(self):
assert (
os.path.exists(get_event_path(self.run_path, kind=V1ArtifactKind.IMAGE))
is False
)
assert (
os.path.exists(get_asset_path(self.run_path, kind=V1ArtifactKind.IMAGE))
is False
)
with patch("polyaxon.tracking.run.Run._log_has_events") as log_dashboard:
self.run.log_image(
name="my_image", data=tensor_np(shape=(1, 8, 8)), dataformats="CHW"
)
assert log_dashboard.call_count == 1
self.event_logger.flush()
assert (
os.path.exists(get_asset_path(self.run_path, kind=V1ArtifactKind.IMAGE))
is True
)
assert (
os.path.exists(get_event_path(self.run_path, kind=V1ArtifactKind.IMAGE))
is True
)
events_file = get_event_path(
self.run_path, kind=V1ArtifactKind.IMAGE, name="my_image"
)
assert os.path.exists(events_file) is True
results = V1Events.read(kind="image", name="my_image", data=events_file)
assert len(results.df.values) == 1
asset_file = get_asset_path(
self.run_path, kind=V1ArtifactKind.IMAGE, name="my_image", ext="png"
)
assert os.path.exists(asset_file) is True
def test_log_artifact(self):
assert (os.path.exists(
get_event_path(self.run_path, kind=V1ArtifactKind.TSV)) is False)
assert (os.path.exists(
get_asset_path(self.run_path, kind=V1ArtifactKind.TSV)) is False)
tsv_file = tempfile.mkdtemp() + "/file.tsv"
self.touch(tsv_file)
self.run.log_artifact(name="file",
path=tsv_file,
artifact_kind=V1ArtifactKind.TSV)
self.event_logger.flush()
assert (os.path.exists(
get_asset_path(self.run_path, kind=V1ArtifactKind.TSV)) is True)
assert (os.path.exists(
get_event_path(self.run_path, kind=V1ArtifactKind.TSV)) is True)
events_file = get_event_path(self.run_path,
kind=V1ArtifactKind.TSV,
name="file")
assert os.path.exists(events_file) is True
results = V1Events.read(kind=V1ArtifactKind.TSV,
name="file",
data=events_file)
assert len(results.df.values) == 1
asset_file = get_asset_path(self.run_path,
kind=V1ArtifactKind.TSV,
name="file",
ext="tsv")
assert os.path.exists(asset_file) is True
def test_log_image_from_path_with_step(self):
assert (
os.path.exists(get_event_path(self.run_path, kind=V1ArtifactKind.IMAGE))
is False
)
assert (
os.path.exists(get_asset_path(self.run_path, kind=V1ArtifactKind.IMAGE))
is False
)
image_file = tempfile.mkdtemp() + "/file.png"
self.touch(image_file)
with patch("polyaxon.tracking.run.Run._log_has_events") as log_dashboard:
self.run.log_image(name="my_image", data=image_file, step=1)
assert log_dashboard.call_count == 1
self.event_logger.flush()
assert (
os.path.exists(get_asset_path(self.run_path, kind=V1ArtifactKind.IMAGE))
is True
)
assert (
os.path.exists(get_event_path(self.run_path, kind=V1ArtifactKind.IMAGE))
is True
)
events_file = get_event_path(
self.run_path, kind=V1ArtifactKind.IMAGE, name="my_image"
)
assert os.path.exists(events_file) is True
results = V1Events.read(kind="image", name="my_image", data=events_file)
assert len(results.df.values) == 1
asset_file = get_asset_path(
self.run_path, kind=V1ArtifactKind.IMAGE, name="my_image", step=1, ext="png"
)
assert os.path.exists(asset_file) is True
def test_log_single_metric(self):
assert (
os.path.exists(get_event_path(self.run_path, kind=V1ArtifactKind.METRIC))
is False
)
assert (
os.path.exists(get_asset_path(self.run_path, kind=V1ArtifactKind.METRIC))
is False
)
with patch("polyaxon.tracking.run.Run._log_has_events") as log_dashboard:
self.run.log_metrics(step=1, metric1=1.1)
assert log_dashboard.call_count == 1
self.event_logger.flush()
assert (
os.path.exists(get_asset_path(self.run_path, kind=V1ArtifactKind.METRIC))
is False
)
assert (
os.path.exists(get_event_path(self.run_path, kind=V1ArtifactKind.METRIC))
is True
)
events_file = get_event_path(
self.run_path, kind=V1ArtifactKind.METRIC, name="metric1"
)
assert os.path.exists(events_file) is True
results = V1Events.read(kind="metric", name="metric1", data=events_file)
assert len(results.df.values) == 1
def test_log_text(self):
assert (
os.path.exists(get_event_path(self.run_path, kind=V1ArtifactKind.TEXT))
is False
)
assert (
os.path.exists(get_asset_path(self.run_path, kind=V1ArtifactKind.TEXT))
is False
)
with patch("polyaxon.tracking.run.Run._log_has_events") as log_dashboard:
self.run.log_text(name="my_text", text="some text", step=1)
assert log_dashboard.call_count == 1
self.event_logger.flush()
assert (
os.path.exists(get_asset_path(self.run_path, kind=V1ArtifactKind.TEXT))
is False
)
assert (
os.path.exists(get_event_path(self.run_path, kind=V1ArtifactKind.TEXT))
is True
)
events_file = get_event_path(
self.run_path, kind=V1ArtifactKind.TEXT, name="my_text"
)
assert os.path.exists(events_file) is True
results = V1Events.read(kind="text", name="my_text", data=events_file)
assert len(results.df.values) == 1
def test_log_mpl_plotly(self):
assert (
os.path.exists(get_event_path(self.run_path, kind=V1ArtifactKind.CHART))
is False
)
assert (
os.path.exists(get_asset_path(self.run_path, kind=V1ArtifactKind.CHART))
is False
)
figure, axes = plt.figure(), plt.gca()
circle1 = plt.Circle((0.2, 0.5), 0.2, color="r")
circle2 = plt.Circle((0.8, 0.5), 0.2, color="g")
axes.add_patch(circle1)
axes.add_patch(circle2)
plt.axis("scaled")
plt.tight_layout()
with patch("polyaxon.tracking.run.Run._log_has_events") as log_dashboard:
self.run.log_mpl_plotly_chart(name="figure", figure=figure, step=1)
assert log_dashboard.call_count == 1
self.event_logger.flush()
assert (
os.path.exists(get_asset_path(self.run_path, kind=V1ArtifactKind.CHART))
is False
)
assert (
os.path.exists(get_event_path(self.run_path, kind=V1ArtifactKind.CHART))
is True
)
events_file = get_event_path(
self.run_path, kind=V1ArtifactKind.CHART, name="figure"
)
assert os.path.exists(events_file) is True
results = V1Events.read(kind="image", name="figure", data=events_file)
assert len(results.df.values) == 1
with patch("polyaxon.tracking.run.Run._log_has_events") as log_dashboard:
self.run.log_mpl_plotly_chart(name="figure", figure=figure, step=2)
assert log_dashboard.call_count == 1
assert plt.fignum_exists(figure.number) is False
self.event_logger.flush()
assert (
os.path.exists(get_asset_path(self.run_path, kind=V1ArtifactKind.CHART))
is False
)
assert (
os.path.exists(get_event_path(self.run_path, kind=V1ArtifactKind.CHART))
is True
)
events_file = get_event_path(
self.run_path, kind=V1ArtifactKind.CHART, name="figure"
)
assert os.path.exists(events_file) is True
results = V1Events.read(kind="image", name="figure", data=events_file)
assert len(results.df.values) == 2
def test_log_mpl_images(self):
assert (os.path.exists(
get_event_path(self.run_path, kind=V1ArtifactKind.IMAGE)) is False)
assert (os.path.exists(
get_asset_path(self.run_path, kind=V1ArtifactKind.IMAGE)) is False)
figures = []
for i in range(5):
figure = plt.figure()
plt.plot([i * 1, i * 2, i * 3], label="Plot " + str(i))
plt.xlabel("X")
plt.xlabel("Y")
plt.legend()
plt.tight_layout()
figures.append(figure)
self.run.log_mpl_image(name="figure",
data=figures,
step=1,
close=False)
assert all(
[plt.fignum_exists(figure.number) is True for figure in figures])
self.event_logger.flush()
assert (os.path.exists(
get_asset_path(self.run_path, kind=V1ArtifactKind.IMAGE)) is True)
assert (os.path.exists(
get_event_path(self.run_path, kind=V1ArtifactKind.IMAGE)) is True)
events_file = get_event_path(self.run_path,
kind=V1ArtifactKind.IMAGE,
name="figure")
assert os.path.exists(events_file) is True
results = V1Events.read(kind="image", name="figure", data=events_file)
assert len(results.df.values) == 1
self.run.log_mpl_image(name="figure", data=figures, step=2)
assert all(
[plt.fignum_exists(figure.number) is False for figure in figures])
self.event_logger.flush()
assert (os.path.exists(
get_asset_path(self.run_path, kind=V1ArtifactKind.IMAGE)) is True)
assert (os.path.exists(
get_event_path(self.run_path, kind=V1ArtifactKind.IMAGE)) is True)
events_file = get_event_path(self.run_path,
kind=V1ArtifactKind.IMAGE,
name="figure")
assert os.path.exists(events_file) is True
results = V1Events.read(kind="image", name="figure", data=events_file)
assert len(results.df.values) == 2
asset_file = get_asset_path(self.run_path,
kind=V1ArtifactKind.IMAGE,
name="figure",
step=1,
ext="png")
assert os.path.exists(asset_file) is True
def test_log_mpl_image(self):
assert (os.path.exists(
get_event_path(self.run_path, kind=V1ArtifactKind.IMAGE)) is False)
assert (os.path.exists(
get_asset_path(self.run_path, kind=V1ArtifactKind.IMAGE)) is False)
figure, axes = plt.figure(), plt.gca()
circle1 = plt.Circle((0.2, 0.5), 0.2, color="r")
circle2 = plt.Circle((0.8, 0.5), 0.2, color="g")
axes.add_patch(circle1)
axes.add_patch(circle2)
plt.axis("scaled")
plt.tight_layout()
self.run.log_mpl_image(name="figure", data=figure, step=1, close=False)
assert plt.fignum_exists(figure.number) is True
self.event_logger.flush()
assert (os.path.exists(
get_asset_path(self.run_path, kind=V1ArtifactKind.IMAGE)) is True)
assert (os.path.exists(
get_event_path(self.run_path, kind=V1ArtifactKind.IMAGE)) is True)
events_file = get_event_path(self.run_path,
kind=V1ArtifactKind.IMAGE,
name="figure")
assert os.path.exists(events_file) is True
results = V1Events.read(kind="image", name="figure", data=events_file)
assert len(results.df.values) == 1
asset_file = get_asset_path(self.run_path,
kind=V1ArtifactKind.IMAGE,
name="figure",
step=1,
ext="png")
assert os.path.exists(asset_file) is True
self.run.log_mpl_image(name="figure", data=figure, step=2)
assert plt.fignum_exists(figure.number) is False
self.event_logger.flush()
assert (os.path.exists(
get_asset_path(self.run_path, kind=V1ArtifactKind.IMAGE)) is True)
assert (os.path.exists(
get_event_path(self.run_path, kind=V1ArtifactKind.IMAGE)) is True)
events_file = get_event_path(self.run_path,
kind=V1ArtifactKind.IMAGE,
name="figure")
assert os.path.exists(events_file) is True
results = V1Events.read(kind="image", name="figure", data=events_file)
assert len(results.df.values) == 2
asset_file = get_asset_path(self.run_path,
kind=V1ArtifactKind.IMAGE,
name="figure",
step=1,
ext="png")
assert os.path.exists(asset_file) is True
def test_log_empty_metric(self):
assert (os.path.exists(
get_event_path(self.run_path, kind=V1ArtifactKind.METRIC)) is
False)
assert (os.path.exists(
get_asset_path(self.run_path, kind=V1ArtifactKind.METRIC)) is
False)
self.run.log_metrics()
self.event_logger.flush()
assert (os.path.exists(
get_event_path(self.run_path, kind=V1ArtifactKind.METRIC)) is
False)
assert (os.path.exists(
get_asset_path(self.run_path, kind=V1ArtifactKind.METRIC)) is
False)
def test_log_charts(self):
x = [1, 2, 3, 4, 5]
y = [6, 7, 2, 4, 5]
bokeh_test = figure(title="simple line example",
x_axis_label="x",
y_axis_label="y")
bokeh_test.line(x, y, line_width=2)
x1 = np.random.randn(200) - 2
x2 = np.random.randn(200)
x3 = np.random.randn(200) + 2
hist_data = [x1, x2, x3]
group_labels = ["Group 1", "Group 2", "Group 3"]
plotly_test = figure_factory.create_distplot(hist_data,
group_labels,
bin_size=[0.1, 0.25, 0.5])
assert (os.path.exists(
get_event_path(self.run_path, kind=V1ArtifactKind.CHART)) is False)
assert (os.path.exists(
get_asset_path(self.run_path, kind=V1ArtifactKind.CHART)) is False)
self.run.log_bokeh_chart(name="bokeh_test", figure=bokeh_test, step=1)
self.run.log_plotly_chart(name="plotly_test",
figure=plotly_test,
step=1)
self.event_logger.flush()
assert (os.path.exists(
get_asset_path(self.run_path, kind=V1ArtifactKind.CHART)) is False)
assert (os.path.exists(
get_event_path(self.run_path, kind=V1ArtifactKind.CHART)) is True)
events_file = get_event_path(self.run_path,
kind=V1ArtifactKind.CHART,
name="bokeh_test")
assert os.path.exists(events_file) is True
results = V1Events.read(kind=V1ArtifactKind.CHART,
name="bokeh_test",
data=events_file)
assert len(results.df.values) == 1
events_file = get_event_path(self.run_path,
kind=V1ArtifactKind.CHART,
name="plotly_test")
assert os.path.exists(events_file) is True
results = V1Events.read(kind=V1ArtifactKind.CHART,
name="plotly_test",
data=events_file)
assert len(results.df.values) == 1
def test_log_curves(self):
x = [1, 2, 3, 4, 5]
y = [6, 7, 2, 4, 5]
with patch("polyaxon.tracking.run.Run._log_has_events") as log_dashboard:
self.run.log_roc_auc_curve(name="roc_test", fpr=x, tpr=y, auc=0.6, step=1)
self.run.log_pr_curve(
name="pr_test", precision=x, recall=y, average_precision=0.6, step=1
)
self.run.log_curve(name="curve_test", x=x, y=y, annotation=0.6, step=1)
assert log_dashboard.call_count == 3
self.event_logger.flush()
assert (
os.path.exists(get_asset_path(self.run_path, kind=V1ArtifactKind.CURVE))
is False
)
assert (
os.path.exists(get_event_path(self.run_path, kind=V1ArtifactKind.CURVE))
is True
)
events_file = get_event_path(
self.run_path, kind=V1ArtifactKind.CURVE, name="roc_test"
)
assert os.path.exists(events_file) is True
results = V1Events.read(
kind=V1ArtifactKind.CURVE, name="roc_test", data=events_file
)
assert len(results.df.values) == 1
events_file = get_event_path(
self.run_path, kind=V1ArtifactKind.CURVE, name="pr_test"
)
assert os.path.exists(events_file) is True
results = V1Events.read(
kind=V1ArtifactKind.CHART, name="pr_test", data=events_file
)
assert len(results.df.values) == 1
events_file = get_event_path(
self.run_path, kind=V1ArtifactKind.CURVE, name="curve_test"
)
assert os.path.exists(events_file) is True
results = V1Events.read(
kind=V1ArtifactKind.CHART, name="curve_test", data=events_file
)
assert len(results.df.values) == 1
def test_log_text(self):
assert (os.path.exists(
get_event_path(self.run_path, kind=V1ArtifactKind.TEXT)) is False)
assert (os.path.exists(
get_asset_path(self.run_path, kind=V1ArtifactKind.TEXT)) is False)
self.run.log_text(name="my_text", text="some text", step=1)
self.event_logger.flush()
assert (os.path.exists(
get_asset_path(self.run_path, kind=V1ArtifactKind.TEXT)) is False)
assert (os.path.exists(
get_event_path(self.run_path, kind=V1ArtifactKind.TEXT)) is True)
events_file = get_event_path(self.run_path,
kind=V1ArtifactKind.TEXT,
name="my_text")
assert os.path.exists(events_file) is True
results = V1Events.read(kind="text", name="my_text", data=events_file)
assert len(results.df.values) == 1
def test_log_html(self):
assert (os.path.exists(
get_event_path(self.run_path, kind=V1ArtifactKind.HTML)) is False)
assert (os.path.exists(
get_asset_path(self.run_path, kind=V1ArtifactKind.HTML)) is False)
self.run.log_html(name="my_div", html="<div>test<div/>", step=1)
self.event_logger.flush()
assert (os.path.exists(
get_asset_path(self.run_path, kind=V1ArtifactKind.HTML)) is False)
assert (os.path.exists(
get_event_path(self.run_path, kind=V1ArtifactKind.HTML)) is True)
events_file = get_event_path(self.run_path,
kind=V1ArtifactKind.HTML,
name="my_div")
assert os.path.exists(events_file) is True
results = V1Events.read(kind="html", name="my_div", data=events_file)
assert len(results.df.values) == 1