def bool_from_text(self, **kwargs):
trial_value = kwargs['trial_value']
sample_key = kwargs['sample_key']
if trial_value.upper() == 'TRUE':
return QueryTransformerResult({sample_key: True}, False)
elif trial_value.upper() == 'FALSE':
return QueryTransformerResult({sample_key: False}, False)
def variant_category_map(self, **kwargs):
trial_value = kwargs['trial_value']
sample_key = kwargs['sample_key']
variant_category_map = {
"Copy Number Variation".lower(): "CNV",
"Any Variation".lower(): {
"$in": ["MUTATION", "CNV"]
},
"Structural Variation".lower(): "SV"
}
trial_value, negate = self.transform.is_negate(trial_value)
# if a curation calls for a Structural Variant, search the free text in the extended_attributes document under
# STRUCTURAL_VARIANT_COMMENT for mention of the TRUE_HUGO_SYMBOL
if trial_value == 'Structural Variation':
sample_value = variant_category_map.get(trial_value.lower())
results = QueryTransformerResult()
results.add_result({'STRUCTURAL_VARIANT_COMMENT': None}, negate)
results.add_result(
{
'STRUCTURED_SV': None,
sample_key: sample_value
}, negate)
return results
elif trial_value.lower() in variant_category_map:
return QueryTransformerResult(
{sample_key: variant_category_map[trial_value.lower()]},
negate)
else:
return QueryTransformerResult({sample_key: trial_value.upper()},
negate)
def external_file_mapping(self, **kwargs):
trial_value = kwargs['trial_value']
sample_key = kwargs['sample_key']
file = kwargs['file']
if file not in self.resources:
with open(self.resource_paths[file]) as file_handle:
self.resources[file] = json.load(file_handle)
resource = self.resources[file]
trial_value, negate = is_negate(trial_value)
match_value = resource.setdefault(trial_value, trial_value)
if match_value.__class__ is list:
return QueryTransformerResult({sample_key: {"$in": sorted(match_value)}}, negate)
else:
return QueryTransformerResult({sample_key: match_value}, negate)
def wildcard_regex(self, **kwargs):
"""
When trial curation criteria include a wildcard prefix (e.g. WILDCARD_PROTEIN_CHANGE), a genomic query must
use a $regex to search for all genomic documents which match the protein prefix.
E.g.
Trial curation match clause:
| genomic:
| wildcard_protein_change: p.R132
Patient genomic data:
| true_protein_change: p.R132H
The above should match in a mongo query.
"""
trial_value = kwargs['trial_value']
# By convention, all protein changes being with "p."
trial_value, negate = self.transform.is_negate(trial_value)
if not trial_value.startswith('p.'):
trial_value = re.escape('p.' + trial_value)
trial_value = f'^{trial_value}[ACDEFGHIKLMNPQRSTVWY]$'
return QueryTransformerResult({kwargs['sample_key']: {'$regex': re.compile(trial_value, re.IGNORECASE)}},
negate)
def nomap(self, **kwargs):
trial_path = kwargs['trial_path']
trial_key = kwargs['trial_key']
trial_value = kwargs['trial_value']
sample_key = kwargs['sample_key']
trial_value, negate = is_negate(trial_value)
return QueryTransformerResult({sample_key: trial_value}, negate)
def allele_fraction_map(self, **kwargs):
operator_map = {
">": "$gt",
"<": "$lt"
}
trial_value = kwargs['trial_value']
trial_value, negate = self.transform.is_negate(trial_value)
sample_key = kwargs['sample_key']
operator = ''.join([i for i in trial_value if not i.isdigit() and i != '.'])
numeric = "".join([i for i in trial_value if i.isdigit() or i == '.'])
return QueryTransformerResult({sample_key: {operator_map[operator]: float(numeric)}}, False)
def cnv_map(self, **kwargs):
# Heterozygous deletion,
# Gain,
# Homozygous deletion,
# High level amplification,
# Neu
trial_value = kwargs['trial_value']
sample_key = kwargs['sample_key']
cnv_map = {
"High Amplification": "High level amplification",
"Homozygous Deletion": "Homozygous deletion",
'Low Amplification': 'Gain',
'Heterozygous Deletion': 'Heterozygous deletion'
}
trial_value, negate = self.transform.is_negate(trial_value)
if trial_value in cnv_map:
return QueryTransformerResult({sample_key: cnv_map[trial_value]}, negate)
else:
return QueryTransformerResult({sample_key: trial_value}, negate)
def mmr_ms_map(self, **kwargs):
mmr_map = {
'MMR-Proficient': 'Proficient (MMR-P / MSS)',
'MMR-Deficient': 'Deficient (MMR-D / MSI-H)',
'MSI-H': 'Deficient (MMR-D / MSI-H)',
'MSI-L': 'Proficient (MMR-P / MSS)',
'MSS': 'Proficient (MMR-P / MSS)'
}
trial_value = kwargs['trial_value']
trial_value, negate = self.transform.is_negate(trial_value)
sample_key = kwargs['sample_key']
sample_value = mmr_map[trial_value]
return QueryTransformerResult({sample_key: sample_value}, negate)
def tmb_range_to_query(self, **kwargs):
sample_key = kwargs['sample_key']
trial_value = kwargs['trial_value']
operator_map = {
"==": "$eq",
"<=": "$lte",
">=": "$gte",
">": "$gt",
"<": "$lt"
}
operator = ''.join([i for i in trial_value if not i.isdigit() and i != '.'])
numeric = "".join([i for i in trial_value if i.isdigit() or i == '.'])
if numeric.startswith('.'):
numeric = '0' + numeric
return QueryTransformerResult({sample_key: {operator_map[operator]: float(numeric)}}, False)
def age_range_to_date_query(self, **kwargs):
sample_key = kwargs['sample_key']
trial_value = kwargs['trial_value']
operator_map = {
"==": "$eq",
"<=": "$gte",
">=": "$lte",
">": "$lte",
"<": "$gte"
}
# funky logic is because 1 month curation is curated as "0.083" (1/12 a year)
operator = ''.join([i for i in trial_value if not i.isdigit() and i != '.'])
numeric = "".join([i for i in trial_value if i.isdigit() or i == '.'])
if numeric.startswith('.'):
numeric = '0' + numeric
split_time = numeric.split('.')
years = int(split_time[0] if split_time[0].isdigit() else 0)
months_fraction = float('0.' + split_time[1]) if len(split_time) > 1 else 0
months = round(months_fraction * 12)
current_date = datetime.date.today()
query_date = current_date + (- relativedelta(years=years, months=months) + relativedelta(hours=6))
query_datetime = datetime.datetime(query_date.year, query_date.month, query_date.day, query_date.hour, 0, 0, 0)
return QueryTransformerResult({sample_key: {operator_map[operator]: query_datetime}}, False)
def to_upper(self, **kwargs):
trial_value = kwargs['trial_value']
sample_key = kwargs['sample_key']
trial_value, negate = is_negate(trial_value)
return QueryTransformerResult({sample_key: trial_value.upper()}, negate)