Skip to content

Commit

Permalink
Recovered the fixes for EI-CoreBioinformatics#124. Testing with TRAVIS
Browse files Browse the repository at this point in the history
  • Loading branch information
@lucventurini
lucventurini committed Nov 1, 2018
1 parent 03ce93f commit 7655819
Show file tree
Hide file tree
Showing 2 changed files with 74 additions and 61 deletions.
2 changes: 1 addition & 1 deletion .travis.yml
View file
Original file line number Diff line number Diff line change
Expand Up @@ -21,8 +21,8 @@ install:
- pip install codecov;
- pip install -r requirements.txt
script:
- touch plants.yaml && python -c "import Mikado; print(Mikado.version)" && rm plants.yaml; # This is to verify we fixed bug #124
- coverage run setup.py nosetests -exe;
- touch plants.yaml && python -c "import Mikado"; # This is to verify we fixed bug #124
after_success:
- codecov

133 changes: 73 additions & 60 deletions Mikado/configuration/configurator.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -23,6 +23,7 @@
from ..exceptions import InvalidJson, UnrecognizedRescaler
from ..utilities import merge_dictionaries
from ..utilities.log_utils import create_default_logger
from typing import Hashable

# from frozendict import frozendict

Expand Down Expand Up @@ -115,6 +116,9 @@ def check_scoring(json_conf):
:param json_conf: configuration dictionary to check.
:type json_conf: dict
:param key: key in the dictionary leading to the store to check. Default: "scoring"
:type key: str
Function to check the "scoring" section of the configuration.
Each scoring function will be checked for:
- validity of the expression (it can be interpreted by Mikado)
Expand All @@ -135,12 +139,15 @@ def check_scoring(json_conf):
invalid_raw = set()
invalid_filter = set()
available_metrics = Transcript.get_available_metrics()
if "scoring" not in json_conf or len(json_conf["scoring"].keys()) == 0:
raise InvalidJson("No parameters specified for scoring!")
if "scoring" not in json_conf:
raise InvalidJson("\"Scoring\" key not find in the scoring dictionary.")
elif len(json_conf["scoring"]) == 0:
raise InvalidJson("Empty \"scoring\" dictionary.")

jdict = json_conf["scoring"].copy()
validator = extend_with_default(jsonschema.Draft4Validator)

for parameter in json_conf["scoring"]:
for parameter in jdict:
if parameter not in available_metrics:
# Leniency for external_scores
if "." in parameter and len(parameter.split(".")) == 2 and parameter.split(".")[0] == "external":
Expand All @@ -152,29 +159,29 @@ def check_scoring(json_conf):
double_parameters.append(parameter)

if not jsonschema.Draft4Validator(scoring_schema).is_valid(
json_conf["scoring"][parameter]):
jdict[parameter]):
errors = [str(_) for _ in list(jsonschema.Draft4Validator(scoring_schema).iter_errors(
json_conf["scoring"][parameter]))]
jdict[parameter]))]
raise InvalidJson("Invalid scoring for {}:\n{}".format(
parameter, "\n".join(errors)))

try:
validator(scoring_schema).validate(json_conf["scoring"][parameter])
validator(scoring_schema).validate(jdict[parameter])
except Exception as err:
raise ValueError(parameter, err)

if ("external" not in parameter and
getattr(Transcript, parameter).usable_raw is False and
json_conf["scoring"][parameter]["use_raw"] is True):
jdict[parameter]["use_raw"] is True):
invalid_raw.add(parameter)

if json_conf["scoring"][parameter]["rescaling"] == "target":
if jdict[parameter]["rescaling"] == "target":
if "value" not in json_conf["scoring"][parameter]:
message = """Target rescaling requested for {0} but no target value specified.
Please specify it with the \"value\" keyword.\n{1}"""
message = message.format(parameter, json_conf["scoring"][parameter])
raise UnrecognizedRescaler(message)
elif json_conf["scoring"][parameter]["use_raw"] is True:
elif jdict[parameter]["use_raw"] is True:
invalid_raw.add(parameter)

if "filter" in json_conf["scoring"][parameter] and "metric" in json_conf["scoring"][parameter]["filter"]:
Expand All @@ -200,7 +207,7 @@ def check_scoring(json_conf):
\t{0}""".format("\n\t".join(list(invalid_raw)))
raise InvalidJson(err_message)

return json_conf
return jdict


def check_all_requirements(json_conf):
Expand All @@ -225,12 +232,8 @@ def check_all_requirements(json_conf):
if "not_fragmentary" not in json_conf:
json_conf["not_fragmentary"] = dict()
json_conf["not_fragmentary"].update(json_conf["requirements"])
# print("Original:", json_conf["not_fragmentary"]["expression"])
# print("Original:", type(json_conf["not_fragmentary"]["expression"]))

try:
# print("Original:", json_conf["not_fragmentary"]["expression"])
# print("Original:", type(json_conf["not_fragmentary"]["expression"]))
json_conf = check_requirements(json_conf,
require_schema,
"not_fragmentary")
Expand Down Expand Up @@ -450,6 +453,15 @@ def create_validator(simple=False):

def _check_scoring_file(json_conf, logger):

"""The purpose of this section is the following:
- check that the scoring file exists somewhere different from the system folder. If it does, check whether it is
a valid file.
- If the local file is broken / wrong / inexistent, use the system one.
*Emit a warning in case there is a local file*
"""


overwritten = False

if json_conf.get("__loaded_scoring", json_conf["pick"]["scoring_file"]) != json_conf["pick"]["scoring_file"]:
Expand All @@ -458,26 +470,56 @@ def _check_scoring_file(json_conf, logger):
overwritten = True
[json_conf.pop(_, None) for _ in ("scoring", "requirements", "as_requirements", "not_fragmentary")]

if os.path.exists(os.path.abspath(json_conf["pick"]["scoring_file"])):
json_conf["pick"]["scoring_file"] = os.path.abspath(
json_conf["pick"]["scoring_file"])
elif os.path.exists(os.path.join(
os.path.dirname(json_conf["filename"]),
json_conf["pick"]["scoring_file"])):
json_conf["pick"]["scoring_file"] = os.path.join(
os.path.dirname(json_conf["filename"]),
json_conf["pick"]["scoring_file"])
elif os.path.exists(
resource_filename(__name__, os.path.join("scoring_files",
json_conf["pick"]["scoring_file"]))):
json_conf["pick"]["scoring_file"] = resource_filename(
__name__,
os.path.join("scoring_files",
json_conf["pick"]["scoring_file"]))
else:
options = [os.path.abspath(json_conf["pick"]["scoring_file"]),
os.path.abspath(os.path.join(os.path.dirname(json_conf["filename"]),
json_conf["pick"]["scoring_file"])),
resource_filename(__name__,
os.path.join("scoring_files",
json_conf["pick"]["scoring_file"]))
]

found = False

for option in options:
if not os.path.exists(option):
continue
if option.endswith(("yaml", "json")):
with open(option) as scoring_file:
if option.endswith("yaml"):
scoring = yaml.load(scoring_file)
else:
scoring = json.load(scoring_file)
if not isinstance(scoring, dict):
continue

try:
scoring = check_all_requirements(scoring)
scoring = check_scoring(scoring)
except InvalidJson:
continue
json_conf = merge_dictionaries(json_conf, scoring)
found = True
break
elif option.endswith(("model", "pickle")):
with open(option, "rb") as forest:
scoring = pickle.load(forest)
if not isinstance(scoring, dict):
continue
elif "scoring" not in scoring:
continue
elif not isinstance(scoring["scoring"], (RandomForestRegressor, RandomForestClassifier)):
continue
del scoring["scoring"]
json_conf = merge_dictionaries(json_conf, scoring)
json_conf = check_all_requirements(json_conf)

if found is False:
raise InvalidJson(
"Scoring file not found: {0}".format(
json_conf["pick"]["scoring_file"]))

json_conf["__loaded_scoring"] = option

return json_conf, overwritten


Expand Down Expand Up @@ -526,37 +568,8 @@ def check_json(json_conf, simple=False, external_dict=None, logger=None):
validator.validate(json_conf)
assert "files" in json_conf["pick"]

overwritten = False

json_conf, overwritten = _check_scoring_file(json_conf, logger)

if json_conf["pick"]["scoring_file"].endswith(("yaml", "json")):
with open(json_conf["pick"]["scoring_file"]) as scoring_file:
if json_conf["pick"]["scoring_file"].endswith("yaml"):
scoring = yaml.load(scoring_file)
else:
scoring = json.load(scoring_file)
assert isinstance(json_conf, dict) and isinstance(scoring, dict),\
(type(json_conf), type(scoring))
json_conf = merge_dictionaries(json_conf, scoring)
json_conf = check_all_requirements(json_conf)
json_conf = check_scoring(json_conf)

elif json_conf["pick"]["scoring_file"].endswith(("model", "pickle")):
with open(json_conf["pick"]["scoring_file"], "rb") as forest:
scoring = pickle.load(forest)
assert isinstance(scoring, dict)
assert "scoring" in scoring and isinstance(scoring["scoring"], (RandomForestRegressor, RandomForestClassifier))
del scoring["scoring"]
json_conf = merge_dictionaries(json_conf, scoring)
json_conf = check_all_requirements(json_conf)
else:
raise InvalidJson(
"Invalid scoring file: {0}".format(
json_conf["pick"]["scoring_file"]))

json_conf["__loaded_scoring"] = json_conf["pick"]["scoring_file"]

if external_dict is not None:
if not isinstance(external_dict, dict):
raise TypeError("Passed an invalid external dictionary, type {}".format(
Expand Down

0 comments on commit 7655819

Please sign in to comment.