def main():
page_file_dir = '/Users/bamana/Documents/InferLink/workspace/memex/memexpython/input/_template_test'
pageManager = PageManager()
files = [f for f in os.listdir(page_file_dir) if os.path.isfile(os.path.join(page_file_dir, f))]
for the_file in files:
if the_file.startswith('.'):
continue
with codecs.open(os.path.join(page_file_dir, the_file), "r", "utf-8") as myfile:
page_str = myfile.read().encode('utf-8')
pageManager.addPage(the_file, page_str)
pageManager.learnStripes()
(list_markup, list_names) = pageManager.learnListMarkups()
rule_set = pageManager.learnAllRules()
(markup, names) = pageManager.rulesToMarkup(rule_set)
for key in markup.keys():
if key in list_markup:
markup[key].update(list_markup[key])
# print json.dumps(markup, sort_keys=True, indent=2, separators=(',', ': '))
rule_set = pageManager.learnRulesFromMarkup(list_markup)
print json.dumps(json.loads(rule_set.toJson()), sort_keys=True, indent=2, separators=(',', ': '))
def induce_template(self, cluster_members):
sub_page_mgr = PageManager()
for id in cluster_members:
curr_page = self.__page_manager.getPage(id)
sub_page_mgr.addPage(id, curr_page.string)
sub_page_mgr.learnStripes()
return sub_page_mgr.getStripes()
def do_learning():
if request.method == 'POST':
data = request.get_json(force=True)
project_folder = data['project_folder']
directory = os.path.join(app.static_folder, 'project_folders', project_folder)
markup_file = os.path.join(directory, 'learning', 'markup.json')
with codecs.open(markup_file, "r", "utf-8") as myfile:
json_str = myfile.read().encode('utf-8')
markup = json.loads(json_str)
pageManager = PageManager()
for key in markup['__URLS__']:
page_file = os.path.join(directory, key)
with codecs.open(page_file, "r", "utf-8") as myfile:
page_str = myfile.read().encode('utf-8')
pageManager.addPage(key, page_str)
markup.pop("__SCHEMA__", None)
markup.pop("__URLS__", None)
pageManager.learnStripes(markup)
rule_set = pageManager.learnRulesFromMarkup(markup)
rules_file = os.path.join(directory, 'learning', 'rules.json')
with codecs.open(rules_file, "w", "utf-8") as myfile:
myfile.write(json.dumps(json.loads(rule_set.toJson()), sort_keys=True, indent=2, separators=(',', ': ')))
myfile.close()
return jsonify(rules = json.loads(rule_set.toJson()) )
abort(404)
def main(argv=None):
if argv is None:
argv = sys.argv
try:
try:
opts, args = getopt.getopt(argv[1:], "dh", ["debug", "help"])
write_debug_files = False
for opt in opts:
if opt in [('-d', ''), ('--debug', '')]:
write_debug_files = True
if opt in [('-h', ''), ('--help', '')]:
raise Usage('python -m learning.RuleLearner [OPTIONAL_PARAMS] [TEST_FILES_FOLDER] [MARKUP_FILE]\n\t[OPTIONAL_PARAMS]: -d to get debug stripe html files')
except getopt.error, msg:
raise Usage(msg)
logger.info('Running RuleLearner with file at %s for rules %s', args[0], args[1])
#read the directory location from arg0
page_file_dir = args[0]
pageManager = PageManager(write_debug_files)
start_time = time.time()
for subdir, dirs, files in os.walk(page_file_dir):
for the_file in files:
if the_file.startswith('.'):
continue
with codecs.open(os.path.join(subdir, the_file), "r", "utf-8") as myfile:
page_str = myfile.read().encode('utf-8')
pageManager.addPage(the_file, page_str)
logger.info("--- LOAD PAGES: %s seconds ---" % (time.time() - start_time))
#Read the markups from a file...
start_time = time.time()
markups_file = args[1]
with codecs.open(markups_file, "r", "utf-8") as myfile:
markup_str = myfile.read().encode('utf-8')
markups = json.loads(markup_str)
markups.pop("__SCHEMA__", None)
markups.pop("__URLS__", None)
logger.info("--- LOAD MARKUPS: %s seconds ---" % (time.time() - start_time))
pageManager.learnStripes(markups)
start_time = time.time()
rule_set = pageManager.learnRulesFromMarkup(markups)
logger.info("--- LEARN RULES FROM MARKUP: %s seconds ---" % (time.time() - start_time))
if(len(args) > 2):
output_file = args[2]
with codecs.open(output_file, "w", "utf-8") as myfile:
myfile.write(rule_set.toJson())
myfile.close()
else:
print rule_set.toJson()
def do_learning():
if request.method == 'POST':
data = request.get_json(force=True)
project_folder = data['project_folder']
directory = os.path.join(app.static_folder, 'project_folders', project_folder)
markup_file = os.path.join(directory, 'learning', 'markup.json')
with codecs.open(markup_file, "r", "utf-8") as myfile:
json_str = myfile.read().encode('utf-8')
markup = json.loads(json_str)
pageManager = PageManager()
for key in markup['__URLS__']:
page_file = os.path.join(directory, key)
with codecs.open(page_file, "r", "utf-8") as myfile:
page_str = myfile.read().encode('utf-8')
pageManager.addPage(key, page_str)
markup.pop("__SCHEMA__", None)
markup.pop("__URLS__", None)
pageManager.learnStripes(markup)
rule_set = pageManager.learnRulesFromMarkup(markup)
rules_file = os.path.join(directory, 'learning', 'rules.json')
with codecs.open(rules_file, "w", "utf-8") as myfile:
myfile.write(json.dumps(json.loads(rule_set.toJson()), sort_keys=True, indent=2, separators=(',', ': ')))
myfile.close()
return jsonify(rules = json.loads(rule_set.toJson()) )
abort(404)
def run(page_file_dir, ignore_files=[]):
test_pages = []
pageManager = PageManager(write_debug_files)
if os.path.isfile(page_file_dir):
with open(page_file_dir) as f:
urls = f.readlines()
for url in urls:
page_url = url.strip()
req = urllib2.urlopen(page_url)
page_contents = req.read()
charset = chardet.detect(page_contents)
page_encoding = charset['encoding']
page_str = page_contents.decode(page_encoding).encode('utf-8')
pageManager.addPage(page_url, page_contents)
test_pages.append(page_url)
else:
files = [
f for f in os.listdir(page_file_dir)
if os.path.isfile(os.path.join(page_file_dir, f))
]
for the_file in files:
if the_file.startswith(
'.'
) or the_file == 'markup.json' or the_file == 'rules.json' or the_file in ignore_files:
continue
with codecs.open(os.path.join(page_file_dir, the_file), "r",
"utf-8") as myfile:
page_str = myfile.read().encode('utf-8')
pageManager.addPage(the_file, page_str)
test_pages.append(the_file)
pageManager.learnStripes()
##table, ul, etc. list learning
# (list_markup, list_names) = pageManager.learnListMarkups()
# list_rules = pageManager.learnRulesFromMarkup(list_markup)
##div learning
# train_pages = {}
# for page_id in pageManager._pages:
# train_pages[page_id] = pageManager.getPage(page_id).getString()
# d = DivListLearner()
# div_rules, div_markup = d.run(train_pages)
rule_set = pageManager.learnAllRules()
rule_set.removeBadRules(test_pages)
# for rule in list_rules.rules:
# rule_set.add_rule(rule)
#
# for rule in div_rules.rules:
# rule_set.add_rule(rule)
return rule_set
def main(argv=None):
if argv is None:
argv = sys.argv
try:
try:
opts, args = getopt.getopt(argv[1:], "h", ["help"])
except getopt.error, msg:
raise Usage(msg)
#read the directory location from arg0
page_file_dir = args[0]
pageManager = PageManager()
page_str_array = []
for subdir, dirs, files in os.walk(page_file_dir):
for the_file in files:
if the_file.startswith('.'):
continue
with codecs.open(os.path.join(subdir, the_file), "r", "utf-8") as myfile:
page_str = myfile.read().encode('utf-8')
page_str_array.append(page_str)
pageManager.addPage(the_file, page_str)
pageManager.learnStripes()
#Read the markups from a file...
markups_file = args[1]
with codecs.open(os.path.join('', markups_file), "r", "utf-8") as myfile:
markup_str = myfile.read().encode('utf-8')
markups = json.loads(markup_str)
markups.pop("__SCHEMA__", None)
#Before we learn the stripes let's make sure we can open the output file
pageManager.learnStripes(markups)
rule_set = pageManager.learnRulesFromMarkup(markups)
if(len(args) > 2):
output_file = args[2]
with codecs.open(output_file, "w", "utf-8") as myfile:
myfile.write(rule_set.toJson())
myfile.close()
#testing
flatten = False
extraction_list = rule_set.extract(page_str_array[0])
if rule_set.validate(extraction_list):
if flatten:
print json.dumps(Landmark.flattenResult(extraction_list), sort_keys=True, indent=2, separators=(',', ': '))
else:
print json.dumps(extraction_list, sort_keys=True, indent=2, separators=(',', ': '))
def learn_list_extractors(self, pages):
page_mgr = PageManager() #write_debug_files=True)
markup = {}
for page in pages:
page_content = pages[page]
page_mgr.addPage(page, page_content)
content_list_markup = self.lists_on_single_page(page_content)
markup[page] = content_list_markup
# print '--- MARKUP ---'
# print json.dumps(markup)
page_mgr.learnStripes(markups=markup)
rules = page_mgr.learnRulesFromMarkup(markup)
# now, for each markup rule, learn a little page manager
sublist_page_managers = {}
for page in markup:
for rule_name in markup[page]:
if rule_name not in sublist_page_managers:
sublist_page_managers[rule_name] = PageManager()
for rid in range(len(markup[page][rule_name]['sequence'])):
row = markup[page][rule_name]['sequence'][rid]
sublist_page_managers[rule_name].addPage(
page + "html%d" % rid, row['extract'])
sublist_sub_rules = {}
for sublist in sublist_page_managers:
sublist_page_managers[sublist].learnStripes()
sub_rules = sublist_page_managers[sublist].learnAllRules(
in_list=True)
sublist_sub_rules[
sublist] = sub_rules # This should match a rule name in the rules...
count = 1
for rule in rules.rules:
# print "== RULE INFO =="
# print str(rule.name)
rule.set_sub_rules(sublist_sub_rules[rule.name])
list_name = '_div_list' + format(count, '04')
for page_id in markup:
if rule.name in markup[page_id]:
markup[page_id][list_name] = markup[page_id].pop(rule.name)
rule.name = list_name
# print str(json.dumps(rule.toJson()))
# print "==============="
#
# print rules.toJson()
return rules, markup
def main(argv=None):
if argv is None:
argv = sys.argv
try:
try:
opts, args = getopt.getopt(argv[1:], "dh", ["debug", "help"])
write_debug_files = False
for opt in opts:
if opt in [('-d', ''), ('--debug', '')]:
write_debug_files = True
if opt in [('-h', ''), ('--help', '')]:
raise Usage('python -m learning.RuleLearnerAllSlots [OPTIONAL_PARAMS] [TEST_FILES_FOLDER] \n\t[OPTIONAL_PARAMS]: -d to get debug stripe html files')
except getopt.error, msg:
raise Usage(msg)
logger.info('Running RuleLearnerAllSlots All Slots with files at %s', args[0])
#read the directory location from arg0
page_file_dir = args[0]
pageManager = PageManager(write_debug_files)
if os.path.isfile(page_file_dir):
with open(page_file_dir) as f:
urls = f.readlines()
for url in urls:
page_url = url.strip()
req = urllib2.urlopen(page_url)
page_contents = req.read()
charset = chardet.detect(page_contents)
page_encoding = charset['encoding']
page_str = page_contents.decode(page_encoding).encode('utf-8')
pageManager.addPage(page_url, page_contents)
else:
files = [f for f in os.listdir(page_file_dir) if os.path.isfile(os.path.join(page_file_dir, f))]
for the_file in files:
if the_file.startswith('.'):
continue
with codecs.open(os.path.join(page_file_dir, the_file), "r", "utf-8") as myfile:
page_str = myfile.read().encode('utf-8')
pageManager.addPage(the_file, page_str)
pageManager.learnStripes()
rule_set = pageManager.learnAllRules()
print json.dumps(json.loads(rule_set.toJson()), sort_keys=True, indent=2, separators=(',', ': '))
def run(page_file_dir, ignore_files = []):
test_pages = []
pageManager = PageManager(write_debug_files)
if os.path.isfile(page_file_dir):
with open(page_file_dir) as f:
urls = f.readlines()
for url in urls:
page_url = url.strip()
req = urllib2.urlopen(page_url)
page_contents = req.read()
charset = chardet.detect(page_contents)
page_encoding = charset['encoding']
page_str = page_contents.decode(page_encoding).encode('utf-8')
pageManager.addPage(page_url, page_contents)
test_pages.append(page_url)
else:
files = [f for f in os.listdir(page_file_dir) if os.path.isfile(os.path.join(page_file_dir, f))]
for the_file in files:
if the_file.startswith('.') or the_file == 'markup.json' or the_file == 'rules.json' or the_file in ignore_files:
continue
with codecs.open(os.path.join(page_file_dir, the_file), "r", "utf-8") as myfile:
page_str = myfile.read().encode('utf-8')
pageManager.addPage(the_file, page_str)
test_pages.append(the_file)
pageManager.learnStripes()
##table, ul, etc. list learning
# (list_markup, list_names) = pageManager.learnListMarkups()
# list_rules = pageManager.learnRulesFromMarkup(list_markup)
##div learning
# train_pages = {}
# for page_id in pageManager._pages:
# train_pages[page_id] = pageManager.getPage(page_id).getString()
# d = DivListLearner()
# div_rules, div_markup = d.run(train_pages)
rule_set = pageManager.learnAllRules()
rule_set.removeBadRules(test_pages)
# for rule in list_rules.rules:
# rule_set.add_rule(rule)
#
# for rule in div_rules.rules:
# rule_set.add_rule(rule)
return rule_set
def learn_list_extractors(self, pages):
page_mgr = PageManager() #write_debug_files=True)
markup = {}
for page in pages:
page_content = pages[page]
page_mgr.addPage(page, page_content)
content_list_markup = self.lists_on_single_page(page_content)
markup[page] = content_list_markup
# print '--- MARKUP ---'
# print json.dumps(markup)
page_mgr.learnStripes(markups=markup)
rules = page_mgr.learnRulesFromMarkup(markup)
# now, for each markup rule, learn a little page manager
sublist_page_managers = {}
for page in markup:
for rule_name in markup[page]:
if rule_name not in sublist_page_managers:
sublist_page_managers[rule_name] = PageManager()
for rid in range(len(markup[page][rule_name]['sequence'])):
row = markup[page][rule_name]['sequence'][rid]
sublist_page_managers[rule_name].addPage(page+"html%d" % rid, row['extract'])
sublist_sub_rules = {}
for sublist in sublist_page_managers:
sublist_page_managers[sublist].learnStripes()
sub_rules = sublist_page_managers[sublist].learnAllRules(in_list = True)
sublist_sub_rules[sublist] = sub_rules # This should match a rule name in the rules...
count = 1
for rule in rules.rules:
# print "== RULE INFO =="
# print str(rule.name)
rule.set_sub_rules(sublist_sub_rules[rule.name])
list_name = '_div_list'+format(count, '04')
for page_id in markup:
if rule.name in markup[page_id]:
markup[page_id][list_name] = markup[page_id].pop(rule.name)
rule.name = list_name
# print str(json.dumps(rule.toJson()))
# print "==============="
#
# print rules.toJson()
return rules, markup
def autolearn_grid():
if request.method == 'POST':
data = request.get_json(force=True)
page_urls = data['urls']
page_manager = PageManager()
results = {}
for page_url in page_urls:
page_contents = urllib2.urlopen(page_url).read()
page_manager.addPage(page_url, page_contents)
page_manager.learnStripes()
rule_set = page_manager.learnAllRules()
results['rules'] = json.loads(rule_set.toJson())
return jsonify(results)
abort(404)
def autolearn_grid():
if request.method == 'POST':
data = request.get_json(force=True)
page_urls = data['urls']
page_manager = PageManager()
results = {}
for page_url in page_urls:
page_contents = urllib2.urlopen(page_url).read()
page_manager.addPage(page_url, page_contents)
page_manager.learnStripes()
rule_set = page_manager.learnAllRules()
results['rules'] = json.loads(rule_set.toJson())
return jsonify(results)
abort(404)
def save_markup():
if request.method == 'POST':
data = request.get_json(force=True)
project_folder = data['project_folder']
markup = data['markup']
directory = os.path.join(app.static_folder, 'project_folders', project_folder)
markup_file = os.path.join(directory, 'learning', 'markup.json')
if not markup['__SCHEMA__'][0]['children']:
markup_slot = {
"id": "j1_2",
"text": "slot",
"icon": "glyphicon glyphicon-stop",
"li_attr": {
"id": "j1_2"
},
"a_attr": {
"href": "#",
"id": "j1_2_anchor"
},
"state": {
"loaded": True,
"opened": False,
"selected": False,
"disabled": False
},
"data": {},
"children": [],
"type": "item"
};
list_slot = {
"a_attr": {
"href": "#",
"id": "j1_3_anchor"
},
"children": [],
"data": {},
"icon": "glyphicon glyphicon-th-list",
"id": "j1_3",
"li_attr": {
"id": "j1_3"
},
"state": {
"disabled": False,
"loaded": True,
"opened": False,
"selected": False
},
"text": "category",
"type": "list"
};
pageManager = PageManager()
for key in markup['__URLS__']:
page_file = os.path.join(directory, key)
with codecs.open(page_file, "r", "utf-8") as myfile:
page_str = myfile.read().encode('utf-8')
pageManager.addPage(key, page_str)
schema = markup.pop("__SCHEMA__", None)
urls = markup.pop("__URLS__", None)
pageManager.learnStripes()
(list_markup, list_names) = pageManager.learnListMarkups()
rule_set = pageManager.learnAllRules()
(markup, names) = pageManager.rulesToMarkup(rule_set)
for key in markup.keys():
if key in list_markup:
markup[key].update(list_markup[key])
count = 1
# Generate the schema from the list slots
for list_name in list_names.keys():
count += 1
auto_markup_slot = copy.deepcopy(list_slot)
auto_markup_slot['text'] = list_name
auto_markup_slot['id'] = 'j1_'+str(count)
auto_markup_slot['li_attr']['id'] = 'j1_'+str(count)
auto_markup_slot['a_attr']['id'] = 'j1_'+str(count)+'_anchor'
## now add the children to the auto learned list slot
children = []
for name in list_names[list_name]:
count += 1
auto_markup_slot_sub = copy.deepcopy(markup_slot)
auto_markup_slot_sub['text'] = name
auto_markup_slot_sub['id'] = 'j1_'+str(count)
auto_markup_slot_sub['li_attr']['id'] = 'j1_'+str(count)
auto_markup_slot_sub['a_attr']['id'] = 'j1_'+str(count)+'_anchor'
children.append(auto_markup_slot_sub)
auto_markup_slot['children'] = children
schema[0]['children'].append(auto_markup_slot)
# Generate the schema from the item slots
for name in names:
count += 1
auto_markup_slot = copy.deepcopy(markup_slot)
auto_markup_slot['text'] = name
auto_markup_slot['id'] = 'j1_'+str(count)
auto_markup_slot['li_attr']['id'] = 'j1_'+str(count)
auto_markup_slot['a_attr']['id'] = 'j1_'+str(count)+'_anchor'
schema[0]['children'].append(auto_markup_slot)
markup['__SCHEMA__'] = schema
markup['__URLS__'] = urls
with codecs.open(markup_file, "w", "utf-8") as myfile:
myfile.write(json.dumps(markup, sort_keys=True, indent=2, separators=(',', ': ')))
myfile.close()
else:
with codecs.open(markup_file, "w", "utf-8") as myfile:
myfile.write(json.dumps(markup, sort_keys=True, indent=2, separators=(',', ': ')))
myfile.close()
return jsonify(markup)
abort(404)
def main(argv=None):
if argv is None:
argv = sys.argv
try:
try:
opts, args = getopt.getopt(argv[1:], "dh", ["debug", "help"])
write_debug_files = False
for opt in opts:
if opt in [('-d', ''), ('--debug', '')]:
write_debug_files = True
if opt in [('-h', ''), ('--help', '')]:
raise Usage(
'python -m learning.RuleLearner [OPTIONAL_PARAMS] [TEST_FILES_FOLDER] [MARKUP_FILE]\n\t[OPTIONAL_PARAMS]: -d to get debug stripe html files'
)
except getopt.error, msg:
raise Usage(msg)
logger.info('Running RuleLearner with file at %s for rules %s',
args[0], args[1])
#read the directory location from arg0
page_file_dir = args[0]
pageManager = PageManager(write_debug_files)
start_time = time.time()
for subdir, dirs, files in os.walk(page_file_dir):
for the_file in files:
if the_file.startswith('.'):
continue
with codecs.open(os.path.join(subdir, the_file), "r",
"utf-8") as myfile:
page_str = myfile.read().encode('utf-8')
pageManager.addPage(the_file, page_str)
logger.info("--- LOAD PAGES: %s seconds ---" %
(time.time() - start_time))
#Read the markups from a file...
start_time = time.time()
markups_file = args[1]
with codecs.open(markups_file, "r", "utf-8") as myfile:
markup_str = myfile.read().encode('utf-8')
markups = json.loads(markup_str)
markups.pop("__SCHEMA__", None)
markups.pop("__URLS__", None)
logger.info("--- LOAD MARKUPS: %s seconds ---" %
(time.time() - start_time))
pageManager.learnStripes(markups)
start_time = time.time()
rule_set = pageManager.learnRulesFromMarkup(markups)
logger.info("--- LEARN RULES FROM MARKUP: %s seconds ---" %
(time.time() - start_time))
if (len(args) > 2):
output_file = args[2]
with codecs.open(output_file, "w", "utf-8") as myfile:
myfile.write(rule_set.toJson())
myfile.close()
else:
print rule_set.toJson()
def save_markup():
if request.method == 'POST':
data = request.get_json(force=True)
project_folder = data['project_folder']
markup = data['markup']
directory = os.path.join(app.static_folder, 'project_folders', project_folder)
markup_file = os.path.join(directory, 'learning', 'markup.json')
if not markup['__SCHEMA__'][0]['children']:
markup_slot = {
"id": "j1_2",
"text": "slot",
"icon": "glyphicon glyphicon-stop",
"li_attr": {
"id": "j1_2"
},
"a_attr": {
"href": "#",
"id": "j1_2_anchor"
},
"state": {
"loaded": True,
"opened": False,
"selected": False,
"disabled": False
},
"data": {},
"children": [],
"type": "item"
};
list_slot = {
"a_attr": {
"href": "#",
"id": "j1_3_anchor"
},
"children": [],
"data": {},
"icon": "glyphicon glyphicon-th-list",
"id": "j1_3",
"li_attr": {
"id": "j1_3"
},
"state": {
"disabled": False,
"loaded": True,
"opened": False,
"selected": False
},
"text": "category",
"type": "list"
};
pageManager = PageManager()
test_pages = []
for key in markup['__URLS__']:
page_file = os.path.join(directory, key)
with codecs.open(page_file, "r", "utf-8") as myfile:
page_str = myfile.read().encode('utf-8')
pageManager.addPage(key, page_str)
test_pages.append(page_str)
schema = markup.pop("__SCHEMA__", None)
urls = markup.pop("__URLS__", None)
pageManager.learnStripes()
list_markup = {}
list_names = {}
if LEARN_LISTS:
(list_markup, list_names) = pageManager.learnListMarkups()
#This is the div learning
train_pages = {}
for page_id in pageManager._pages:
train_pages[page_id] = pageManager.getPage(page_id).getString()
d = DivListLearner()
div_rules, div_markup = d.run(train_pages)
(div_list_markup, div_list_names) = pageManager.listRulesToMarkup(div_rules)
for page_id in div_markup:
for item in div_markup[page_id]:
if item in div_list_markup[page_id]:
if 'starting_token_location' in div_markup[page_id][item]:
div_list_markup[page_id][item]['starting_token_location'] = div_markup[page_id][item]['starting_token_location']
if 'ending_token_location' in div_markup[page_id][item]:
div_list_markup[page_id][item]['ending_token_location'] = div_markup[page_id][item]['ending_token_location']
if div_markup[page_id][item]['sequence']:
for idx, val in enumerate(div_markup[page_id][item]['sequence']):
if len(div_list_markup[page_id][item]['sequence']) <= idx:
div_list_markup[page_id][item]['sequence'].insert(idx, val);
else:
div_list_markup[page_id][item]['sequence'][idx]['starting_token_location'] = val['starting_token_location']
div_list_markup[page_id][item]['sequence'][idx]['ending_token_location'] = val['ending_token_location']
#Now add these to the list_markup and list_names
if len(div_rules.rules) > 0:
for page_id in div_list_markup:
if page_id not in list_markup:
list_markup[page_id] = {}
list_markup[page_id].update(div_list_markup[page_id])
list_names.update(div_list_names)
rule_set = pageManager.learnAllRules()
rule_set.removeBadRules(test_pages)
(markup, names) = pageManager.rulesToMarkup(rule_set)
for key in markup.keys():
if key in list_markup:
markup[key].update(list_markup[key])
count = 1
# Generate the schema from the list slots
for list_name in list_names.keys():
count += 1
auto_markup_slot = copy.deepcopy(list_slot)
auto_markup_slot['text'] = list_name
auto_markup_slot['id'] = 'j1_'+str(count)
auto_markup_slot['li_attr']['id'] = 'j1_'+str(count)
auto_markup_slot['a_attr']['id'] = 'j1_'+str(count)+'_anchor'
## now add the children to the auto learned list slot
children = []
for name in list_names[list_name]:
count += 1
auto_markup_slot_sub = copy.deepcopy(markup_slot)
auto_markup_slot_sub['text'] = name
auto_markup_slot_sub['id'] = 'j1_'+str(count)
auto_markup_slot_sub['li_attr']['id'] = 'j1_'+str(count)
auto_markup_slot_sub['a_attr']['id'] = 'j1_'+str(count)+'_anchor'
children.append(auto_markup_slot_sub)
auto_markup_slot['children'] = children
schema[0]['children'].append(auto_markup_slot)
# Generate the schema from the item slots
for name in names:
count += 1
auto_markup_slot = copy.deepcopy(markup_slot)
auto_markup_slot['text'] = name
auto_markup_slot['id'] = 'j1_'+str(count)
auto_markup_slot['li_attr']['id'] = 'j1_'+str(count)
auto_markup_slot['a_attr']['id'] = 'j1_'+str(count)+'_anchor'
schema[0]['children'].append(auto_markup_slot)
markup['__SCHEMA__'] = schema
markup['__URLS__'] = urls
with codecs.open(markup_file, "w", "utf-8") as myfile:
myfile.write(json.dumps(markup, sort_keys=True, indent=2, separators=(',', ': ')))
myfile.close()
else:
with codecs.open(markup_file, "w", "utf-8") as myfile:
myfile.write(json.dumps(markup, sort_keys=True, indent=2, separators=(',', ': ')))
myfile.close()
return jsonify(markup)
abort(404)
def main(argv=None):
if argv is None:
argv = sys.argv
try:
try:
opts, args = getopt.getopt(argv[1:], "h", ["help"])
except getopt.error, msg:
raise Usage(msg)
#read the directory location from arg0
page_file_dir = args[0]
pageManager = PageManager()
page_str_array = []
for subdir, dirs, files in os.walk(page_file_dir):
for the_file in files:
if the_file.startswith('.'):
continue
with codecs.open(os.path.join(subdir, the_file), "r",
"utf-8") as myfile:
page_str = myfile.read().encode('utf-8')
page_str_array.append(page_str)
pageManager.addPage(the_file, page_str)
pageManager.learnStripes()
#Read the markups from a file...
markups_file = args[1]
with codecs.open(os.path.join('', markups_file), "r",
"utf-8") as myfile:
markup_str = myfile.read().encode('utf-8')
markups = json.loads(markup_str)
markups.pop("__SCHEMA__", None)
#Before we learn the stripes let's make sure we can open the output file
pageManager.learnStripes(markups)
rule_set = pageManager.learnRulesFromMarkup(markups)
if (len(args) > 2):
output_file = args[2]
with codecs.open(output_file, "w", "utf-8") as myfile:
myfile.write(rule_set.toJson())
myfile.close()
#testing
flatten = False
extraction_list = rule_set.extract(page_str_array[0])
if rule_set.validate(extraction_list):
if flatten:
print json.dumps(Landmark.flattenResult(extraction_list),
sort_keys=True,
indent=2,
separators=(',', ': '))
else:
print json.dumps(extraction_list,
sort_keys=True,
indent=2,
separators=(',', ': '))