def get_all_projects():
try:
projects = [
project.to_dict(show=["datasets"])
for project in Project.query.all()
]
return jsonify(projects)
except Exception as e:
return handle_exception(e)
def get_data(rid: str):
try:
record = DatasetRecord.query.filter_by(id=rid).first()
pid = record.project_id
file = BytesIO(record.data)
data = pd.read_parquet(file)
col_infos = DatasetMeta.query.filter_by(project_id=pid).all()
column_info = {}
columns = []
categorical_columns = []
numeric_columns = []
label_column = None
for c in col_infos:
columns.append(c.key)
c_type = c.data_type
if c_type == "label":
label_column = c.key
elif c_type == "numeric":
numeric_columns.append(c.key)
elif c_type == "categorical":
categorical_columns.append(c.key)
col = c.to_dict()
column_info[c.key] = col
d = data.T.to_json()
return (
jsonify({
"values": list(json.loads(d).values()),
"columnInfo": column_info,
"numericColumns": numeric_columns,
"categoricalColumns": categorical_columns,
"labelColumn": label_column,
"columns": columns,
"id": rid,
"version": record.version,
}),
200,
)
except Exception as e:
raise e
return handle_exception(e)
def compute_state():
try:
body = request.get_json()
if body is None:
raise Exception("Resources incorrect!")
provenance = body["provenance"]
data = pd.DataFrame(body["data"])
graph = Graph(**provenance)
return jsonify(
json.loads(json.dumps(graph.states(data), default=lambda o: o.toJSON()))
)
except Exception as ex:
raise ex
return handle_exception(ex)
def create_project():
try:
body = request.get_json()
if body is None:
raise Exception("No project name specified")
p_name = body["p_name"]
project = Project(name=p_name)
db.session.add(project)
db.session.commit()
project = project.to_dict()
return jsonify(project), 200
except Exception as e:
return handle_exception(e)
def compute_compare():
try:
body = request.get_json()
if body is None:
raise Exception("Resources incorrect!")
provenance = body["provenance"]
base = pd.DataFrame(body["base"])
updated = pd.DataFrame(body["target"])
graph = Graph(**provenance)
compare = graph.compare(base, updated)
return jsonify(compare)
except Exception as ex:
raise ex
return handle_exception(ex)
def get_members_route():
try:
body = request.get_json()
if body is None:
raise Exception("No intent provided")
data = body["data"]
intent = body["intent"]
algorithm = body["algorithm"]
params = body["params"]
dimensions = body["dimensions"]
info = body["info"]
data = pd.DataFrame(data)
return jsonify(
get_members(data, dimensions, intent, algorithm, params, info))
except Exception as e:
return handle_exception(e)
def get_predictions():
try:
body = request.get_json()
if body is None:
raise Exception("No interactions found")
data = pd.DataFrame(body["data"])
dimensions = body["dimensions"]
selections = body["selections"]
if len(selections) == 0:
return jsonify([])
inference = Inference(data, selections, dimensions, [])
predictions = inference.predict()
return jsonify(predictions)
except Exception as e:
raise e
return handle_exception(e)
def add_category_column():
try:
body = request.get_json()
if body is None:
raise Exception("No data specified")
pid = body["pid"]
column_name = body["c_name"]
options = body["options"]
letters = string.ascii_uppercase
shorts = [
c.short for c in DatasetMeta.query.filter_by(project_id=pid).all()
]
shorts = list(sorted(shorts))
last_short = shorts[-1]
next_short = None
if last_short == "Z":
raise Exception("Too many categories")
else:
next_short = letters[letters.index(last_short) + 1]
new_category_column = DatasetMeta(
project_id=pid,
key=get_column_key(column_name),
fullname=column_name,
short=next_short,
data_type="categorical",
options=options,
)
db.session.add(new_category_column)
db.session.commit()
col_infos = DatasetMeta.query.filter_by(project_id=pid).all()
column_info = {}
columns = []
categorical_columns = []
numeric_columns = []
label_column = None
for c in col_infos:
columns.append(c.key)
c_type = c.data_type
if c_type == "label":
label_column = c.key
elif c_type == "numeric":
numeric_columns.append(c.key)
elif c_type == "categorical":
categorical_columns.append(c.key)
col = c.to_dict()
column_info[c.key] = col
return jsonify({
"column_info": column_info,
"numeric_columns": numeric_columns,
"categorical_columns": categorical_columns,
"label_column": label_column,
"columns": columns,
})
except Exception as e:
return handle_exception(e)