#!/usr/bin/python
# -*- coding: utf-8 -*-
import cgitb
cgitb.enable()
import sys, os, cgi, json
from HitEx.matching_set import MatchingSet
#activate_this = os.path.join(os.path.dirname(__file__), 'venv/bin/activate_this.py')
#execfile(activate_this, dict(__file__=activate_this))
def create_mset_json(parameters, criteria):
ms = MatchingSet(parameters, criteria)
ms.create_set()
url = ms.get_url()
j = ms.to_json()
return (ms, j, url)
def main():
"""Main CGI handler."""
# Print JSON header
print_header()
# Convert form fields to regular dictionary
form = dict((field, cgi.FieldStorage().getvalue(field)) for field in cgi.FieldStorage().keys())
#form = {"command":"hitex", "query_type":"lemma", "query_w":"hund", "use_defaults":"1"} # for testing
parameters = {}
criteria = {}
default_parameters = {
#"query_type": "lemma", # wordform or cqp or lemma
#"query_w" : u'bröd', # u"huset", [deprel = "SS" & lemma contains "språk"].decode("utf-8")
#"query_pos" : "", # "NN",
#"corpus_list":["ROM99","GP2012","LASBART"], #randomly pick one of them? "BLOGGMIX2012"
"corpus_list":["rom99","flashback-resor","gp2013","gp2d","attasidor","lasbart","suc3","talbanken"], #"wikipedia-sv" "sweachum","sweacsam"
#"corpus_list":["attasidor","lasbart","talbanken","rom99",
# "familjeliv-allmanna-fritid","flashback-mat",
# "flashback-resor"],
"max_kwics": 100, #300, # nr KWICs to process (limited for efficiency reasons)
"maxhit": 10, # maximum number of matches to return
"target_edge" : "end", # to which edge the keyword should be close to
"proportion" : 50, # within which percentage of the sent the keyword should appear
"target_cefr" : "B1", # 'any' not supported - omit readability to obtain the same effect
"voc_thr" : 0, # percentage of words above the target CEFR level
"min_len" : 6,
"max_len" : 20,
"non_alpha_thr": 30,
"non_lemmatized_thr": 30,
"lex_to_func_thr": "", #0.8,
"sensitive_voc_cats": ["all"], # ["sex", "violence", "other", "religion", "secretion"],
"preserve_bad":True}
default_criteria = {
"well_formedness":{"root":"filter", "sent_tokenization":"filter",
"elliptic":"filter", "non_alpha":"filter",
"non_lemmatized":"filter"},
"isolability":{"struct_conn":"filter", "yn_answer":"filter",
"anaphora-PN":"filter", "anaphora-AB":"filter"},
"readability":"filter",
"typicality": "ranker",
"sensitive_voc": "filter",
"other_criteria":{"length":"filter",
"proper_name":"ranker",
"repkw":"filter",
"kw_position":"",
"modal_verb":"",
"participle":"", # same as korp
"sverb":"", # same as korp
"interrogative":"filter",
"neg_form":"",
"abbrev":"filter",
"direct_speech":"filter",
"diff_voc_kelly":"filter",
"svalex_fr":"",
"out_of_svalex":"filter"}
}
#default_criteria = {"readability":"ranker"}
# TO DO: change to {"filters":["readability"], "rankers":[]}
# Get input values
if form.get("command", "") == "hitex":
if "query_type" in form:
parameters["query_type"] = form["query_type"]
else:
json.dumps({"Error":"Missing 'query_type', choose what TYPE of \
term to search for (lemma, wordform or cqp).",
"code":400})
if "query_w" in form:
parameters["query_w"] = form["query_w"]
else:
json.dumps({"Error":"Missing 'query_w', choose a what term to \
search for.", "code":400})
#if "query_pos" in form: #when not specified, any POS will be matched
# parameters["query_pos"] = form["query_pos"]
if "target_cefr" in form:
if form["target_cefr"] not in ["A1", "A2", "B1", "B2", "C1"]:
json.dumps({"Error":"Argument 'target_cefr' has to be a CEFR \
level (A1, A2, B1, B2 or C1)", "code":400})
if "readability" in form and not form.has_key("target_cefr"):
json.dumps({"Error":"To use 'readability' as a criteria, specify \
'target_cefr' (A1, A2, B1, B2 or C1)", "code":400})
if "random_seed" in form:
try:
parameters["random_seed"] = int(form["random_seed"])
except ValueError:
json.dumps({"Error":"The argument 'random_seed' has to be a integer value", "code":400})
#default_parameters["target_cefr"] = form.get("target_cefr", "any")
if form.get("use_defaults") in ["true", True, 1, "1"]:
parameters.update(default_parameters)
criteria.update(default_criteria)
elif "use_defaults" in form and form["use_defaults"] != "true":
json.dumps({"Error":"'use_defaults' should be set to 'true' for \
using default parameters and criteria for the search.",
"code":400})
else:
mandatory_params = []
for p in default_parameters:
if p in form:
parameters[p] = form[p]
elif p not in form and p in mandatory_params:
json.dumps({"Error":"The arguments 'corpus_list', 'maxhit' \
and 'max_kwics' are mandatory. Set 'use_defaults' to 'true' \
if you don't want to specify these.",
"code":400})
for cr in default_criteria:
if type(default_criteria[cr]) == dict:
for sub_cr in default_criteria[cr]:
if sub_cr in form:
parameters[sub_cr] = form[sub_cr]
elif cr in form:
parameters[cr] = form[cr]
# Analyze sentences for match
mset, json_ms, url = create_mset_json(parameters, criteria)
result = json_ms
else:
result = json.dumps({"Error":"Unknown or missing command. Command \
option(s): hitex, complexity, log","code":404})
print_object(json.loads(result), form)
def print_header():
"""Prints the JSON header."""
print "Content-Type: application/json"
print "Access-Control-Allow-Origin: *"
print "Access-Control-Allow-Methods: POST, GET, OPTIONS"
print "Access-Control-Allow-Headers: Content-Type"
print
def print_object(obj, form):
"""Prints an object in JSON format.
The CGI form can contain optional parameters 'callback' and 'indent'
which change the output format.
"""
callback = form.get("callback")
if callback: print callback + "(",
try:
indent = int(form.get("indent"))
print json.dumps(obj, sort_keys=True, indent=indent),
except:
print json.dumps(obj, separators=(",",":"))
if callback: print ")",
print
if __name__ == "__main__":
main()