class StepTwo(Transform):
bar = InputTag("bar")
bongo = InputTag("bongo")
baz = OutputTag("baz")
def script(self):
pass
class StepThree(Transform):
baz = InputTag("baz")
bingo = InputTag("bingo")
bop = OutputTag("bop")
def script(self):
pass
class BasicStep(Transform):
foo = InputTag("foo")
bar = InputTag("bar")
baz = OutputTag("baz")
def script(self):
pass
class StepThree(Transform):
baz = InputTag("baz")
bleep = InputTag("bleep")
boppo = OutputTag("boppo")
def script(self):
with self.baz.open() as f:
data = f.read()
with self.bleep.open() as f:
data2 = f.read()
with self.boppo.open() as f:
f.write(data + " " + data2)
class Predict(Transform):
to_predict_on = InputTag("to_predict_on")
model = InputTag("model")
predictions = OutputTag("predictions")
def script(self):
with self.model.openbin() as f:
model = pickle.load(f)
with self.to_predict_on.openbin() as f:
X = pd.read_pickle(f)
predictions = model.predict(X)
with self.predictions.openbin() as f:
pickle.dump(predictions, f)
class BasicStep(Transform):
foo = InputTag("foo")
bar = InputTag("bar")
baz = OutputTag("baz")
def script(self):
with self.foo.open() as f:
assert f.read() == "foo contents"
with self.bar.open() as f:
assert f.read() == "bar contents"
with self.baz.open() as f:
f.write("baz contents")
class StepOne(Transform):
foo = InputTag("foo")
bar = OutputTag("bar")
bingo = OutputTag("bingo")
def script(self):
pass
class Two(Transform):
foo = InputTag("foo")
baz = OutputTag("baz")
def script(self):
self.foo.touch()
pass
class Train(Transform):
train_df = InputTag("train_df")
train_labels = InputTag("train_labels")
model = OutputTag("model")
def script(self):
with self.train_df.openbin() as f:
X = pd.read_pickle(f)
with self.train_labels.openbin() as f:
y = pd.read_pickle(f)
# Create a classifier: a support vector classifier
classifier = svm.SVC(gamma=0.001)
model = classifier.fit(X, y.values.ravel())
with self.model.openbin() as f:
pickle.dump(model, f)
class StepTwo(Transform):
bar = InputTag("bar")
baz = OutputTag("baz")
def script(self):
with self.bar.open() as f:
data = f.read()
with self.baz.open() as f:
f.write(data)
class StepOne(Transform):
foo = InputTag("foo")
bar = OutputTag("bar")
def script(self):
with self.foo.open() as f:
data = f.read()
with self.bar.open() as f:
f.write(data)
class BasicStep(Transform):
foo = InputTag("foo")
baz = OutputTag("baz")
def script(self):
self.foo.touch()
self.baz.touch()
assert self.foo.ref.opened == True
assert self.baz.ref.opened == True
class Concat(Transform):
source = InputTag("source")
copied = OutputTag("fileset::copied")
def script(self):
with self.source.open() as f:
content = f.read()
for ref in self.copied:
with ref.open() as f:
f.write(content)
class EvaluateResults(Transform):
predictions = InputTag("predictions")
test_labels = InputTag("test_labels")
accuracy = OutputTag("accuracy")
def script(self):
with self.predictions.openbin() as f:
pred = pd.read_pickle(f)
with self.test_labels.openbin() as f:
real = list(pd.read_pickle(f)["label"])
total = len(pred)
correct = 0
for i in range(total):
if pred[i] == real[i]:
correct = correct + 1
accuracy = round(correct * 1.0 / total, 4)
with self.accuracy.open() as f:
f.write(str(accuracy))
class BuildDf(Transform):
raw_images = InputTag("raw_images")
df = OutputTag("df")
labels = OutputTag("labels")
def script(self):
with self.raw_images.open() as f:
data = np.genfromtxt(f, delimiter=",")
labels = data[:, -1].astype(int)
images = data[:, 0:-1] / 16.0
df = pd.DataFrame(images)
labels_df = pd.DataFrame(data={"label": list(labels)})
with self.df.openbin() as f:
df.to_pickle(f)
with self.labels.openbin() as f:
labels_df.to_pickle(f)
from tinybaker import InputTag, OutputTag, cli
infile = InputTag("infile")
outfile = OutputTag("outfile")
def script():
with infile.open() as f:
contents = f.read()
with outfile.open() as f:
f.write(contents + " but different")
if __name__ == "__main__":
cli(locals())
class One(Transform):
foo = InputTag("foo")
baz = OutputTag("baz")
def script(self):
pass
class StepFour(Transform):
bop = InputTag("bop")
bip = OutputTag("bip")
def script(self):
pass
