def task_5(medical=get_medical_terms()):
"""Task 5: Exchange evaluations"""
relevant = lambda chapter: set(chapter.vector.keys()) & medical
with open('etc/groups.json', 'r') as f:
groups = json.load(f)
values = groups['Group 14']
for case_code, results in sorted(values.items(), key=itemgetter(0)):
filtered = [Therapy.ALL[i] for i in results if i in Therapy.ALL]
_precision([(c, None, None, relevant(c)) for c in filtered])
def _task_6_eval(case, medical=get_medical_terms()):
""""Task 6 evaluation."""
relevant = lambda chapter: set(chapter.vector.keys()) & medical
#result_3 = [(c,) for s, c in task_3(case, 1000)]
result_6a = [(c, s, None, relevant(c)) for s, c in task_6a(case, 1000)]
result_6b = [(c, s, None, relevant(c)) for s, c in task_6b(case, 1000)]
#_precision(result_6a)
#_precision(result_6b)
print("[%s]: Kendal Tau: %.3f" % (case.code,
_kendall_tau(result_6a, result_6b, 1000)))
def calculate_case_statistics():
"""Calculate statistics of patient cases."""
words = data.get_stopwords()
terms = data.get_medical_terms()
print("Case | Lines | Stopwords | Terms | Medical terms")
for code, case in sorted(PatientCase.ALL.items()):
print(' & '.join((code, str(len(case.text.split('\n'))),
str(len([i for i in case.text.split() if i in words])),
str(len(case.vector)),
str(len([i for i in case.vector.keys() if i in terms])))) + r' \\')
print()
def task_4(case, medical=get_medical_terms()):
"""Task 4: Evaluate results from task 3."""
#if not hasattr(PatientCase.ALL["1"], 'vector2'):
# from index import create_vectors, _tf_raw_freq, _idf_prob
# create_vectors(idf=_idf_prob, attr='vector2') # B
# #create_vectors(tf=_tf_raw_freq, attr='vector2') # C
# #create_vectors(tf=_tf_raw_freq, idf=_idf_prob, attr='vector3') # D
results = _task_4_search(case, medical)
#results2 = _task_4_search(case, medical, 'vector3')
#print("[%s]: Kendal Tau: %.3f" % (case.code, _kendall_tau(results, results2, 1000)))
#_task_4_print_terms(results)
_precision(results)
def _task_4_print_terms(results, medical=get_medical_terms()):
"""Prints out terms/medical terms etc for task 4"""
print("Rank | Chapter | Score | Relevant | Terms")
for i, tmp in enumerate(results[:10]):
obj, v, terms, rel = tmp
r = 'Yes' if rel else 'No'
b_terms = []
for t in terms:
b_terms.append('\\textbf{%s}' % t if t in medical else t)
#print('\t%i & %s & %.4f & %s & %s \\\\' % (
# i+1, obj.code, v, r, ', '.join(b_terms)))
print('%i | %s | %.4f | %s | %s' % (
i+1, obj.code, v, r, ', '.join(terms)))
def main(script):
"""Run all the functions in this module."""
data.main() # Populate all objects
# Generate a LaTeX table with all stopwords
_generate_columned_table(sorted(data.get_stopwords()),
6, 'stopwords', 'Norwegian stopwords')
# Generate a LaTeX table with all medical terms
_generate_columned_table(sorted(data.get_medical_terms()),
3, 'medicalterms', 'Medical terms')
generate_cases_table()
calculate_chapter_statistics()
calculate_case_statistics()