def load_annotations(body):
"""Create slybot annotations from annotated html."""
if not body:
return {'annotations-plugin': {'extracts': []}}
sel = Selector(text=add_tagids(body))
existing_ids = set()
annotations = []
for elem in sel.xpath('//*[@data-scrapy-annotate]'):
attributes = elem.root.attrib
try:
# Load annotation json and skip malformed json strings
annotation = json.loads(unquote(
attributes['data-scrapy-annotate']))
except ValueError:
continue
if (isinstance(elem.root, _Element)
and elem.root.tag.lower() == 'ins'):
annotation.update(find_generated_annotation(elem))
else:
annotation['tagid'] = attributes.get('data-tagid')
if 'id' not in annotation:
annotation['id'] = gen_id(disallow=existing_ids)
existing_ids.add(annotation['id'])
annotations.append(annotation)
for elem in sel.xpath('//*[@%s]' % '|@'.join(IGNORE_ATTRIBUTES)):
attributes = elem.root.attrib
for attribute in IGNORE_ATTRIBUTES:
if attribute in attributes:
break
ignore = {attribute[len('data-scrapy-'):]: True}
if 'id' not in ignore:
ignore['id'] = gen_id(disallow=existing_ids)
existing_ids.add(ignore['id'])
annotations.append(ignore)
return {'annotations-plugin': {'extracts': annotations}}
def load_annotations(body):
"""Create slybot annotations from annotated html."""
if not body:
return {'annotations-plugin': {'extracts': []}}
sel = Selector(text=add_tagids(body))
existing_ids = set()
annotations = []
for elem in sel.xpath('//*[@data-scrapy-annotate]'):
attributes = elem.root.attrib
try:
# Load annotation json and skip malformed json strings
annotation = json.loads(
unquote(attributes['data-scrapy-annotate']))
except ValueError:
continue
if (isinstance(elem.root, _Element) and
elem.root.tag.lower() == 'ins'):
annotation.update(find_generated_annotation(elem))
else:
annotation['tagid'] = attributes.get('data-tagid')
if 'id' not in annotation:
annotation['id'] = gen_id(disallow=existing_ids)
existing_ids.add(annotation['id'])
annotations.append(annotation)
for elem in sel.xpath('//*[@%s]' % '|@'.join(IGNORE_ATTRIBUTES)):
attributes = elem.root.attrib
for attribute in IGNORE_ATTRIBUTES:
if attribute in attributes:
break
ignore = {attribute[len('data-scrapy-'):]: True}
if 'id' not in ignore:
ignore['id'] = gen_id(disallow=existing_ids)
existing_ids.add(ignore['id'])
annotations.append(ignore)
return {'annotations-plugin': {'extracts': annotations}}
def html4annotation(htmlpage, baseurl=None, proxy_resources=None):
"""Convert the given html document for the annotation UI
This adds tags, removes scripts and optionally adds a base url
"""
htmlpage = add_tagids(htmlpage)
cleaned_html = descriptify(htmlpage, baseurl, proxy=proxy_resources)
return cleaned_html
def html4annotation(htmlpage, baseurl=None, proxy_resources=None):
"""Convert the given html document for the annotation UI
This adds tags, removes scripts and optionally adds a base url
"""
htmlpage = add_tagids(htmlpage)
cleaned_html = descriptify(htmlpage, baseurl, proxy=proxy_resources)
return cleaned_html
def apply_annotations(annotations, target_page, legacy=False):
selector_annotations, tagid_annotations = _filter_annotations(annotations)
inserts = defaultdict(list)
numbered_html = add_tagids(target_page)
if selector_annotations:
converted_annotations = apply_selector_annotations(
selector_annotations, numbered_html)
tagid_annotations += converted_annotations
target = iter(parse_html(numbered_html))
output, tag_stack = [], []
element = next(target)
last_id = 0
# XXX: A dummy element is added to the end so if the last annotation is
# generated it will be added to the output
filtered = defaultdict(list)
for grouped in tagid_annotations:
for ann in arg_to_iter(grouped):
filtered[ann['tagid']].append(ann)
dummy = [(1e9, [{}])]
sorted_annotations = sorted([(int(k), v) for k, v in filtered.items()] +
dummy)
try:
for aid, annotation_data in sorted_annotations:
# Move target until replacement/insertion point
while True:
while not isinstance(element, HtmlTag) or element.tag == 'ins':
output.append(numbered_html[element.start:element.end])
element = next(target)
if element.tag_type in {OPEN_TAG, UNPAIRED_TAG}:
last_id = element.attributes.get(TAGID)
tag_stack.append(last_id)
if element.tag_type in {CLOSE_TAG, UNPAIRED_TAG} and tag_stack:
if ('__added' not in element.attributes and
last_id is not None and aid is not None and
int(last_id) < int(aid)):
output.append(numbered_html[element.start:element.end])
element.attributes['__added'] = True
last_inserted = tag_stack.pop()
to_insert = inserts.pop(last_inserted, None)
if to_insert:
output.extend(to_insert)
# Skip all nodes up to the next HtmlTag as these
# have already been added
while True:
element = next(target)
try:
last_id = element.attributes.get(TAGID,
last_id)
except AttributeError:
pass
if isinstance(element, HtmlTag):
break
continue
if (last_id is not None and aid is not None and
int(last_id) < int(aid)):
if '__added' not in element.attributes:
output.append(numbered_html[element.start:element.end])
element.attributes['__added'] = True
element = next(target)
else:
break
generated = []
next_generated = []
regular_annotations = []
# Place generated annotations at the end and sort by slice
for annotation in sorted(annotation_data, key=_annotation_key):
if annotation.get('generated'):
if annotation.get('insert_after'):
next_generated.append(annotation)
else:
generated.append(annotation)
else:
regular_annotations.append(annotation)
# Add annotations data as required
if regular_annotations:
annotation_info = _gen_annotation_info(regular_annotations,
legacy)
for key, val in annotation_info.items():
element.attributes[key] = val
next_text_section = ''
if generated:
inner_data, target = tee(target)
nodes = _get_inner_nodes(inner_data)
next_text_section = _get_generated_annotation(
element, generated, nodes, numbered_html, inserts,
legacy)
if next_generated:
inner_data, target = tee(target)
open_tags = 0 if element.tag_type == UNPAIRED_TAG else 1
nodes = _get_inner_nodes(inner_data, open_tags=open_tags,
insert_after=True)
next_text_section = _get_generated_annotation(
element, next_generated, nodes, numbered_html, inserts,
legacy)
if '__added' not in element.attributes:
output.append(serialize_tag(element))
element.attributes['__added'] = True
# If an <ins> tag has been inserted we need to move forward
if next_text_section:
while True:
elem = next(target)
if (isinstance(elem, HtmlDataFragment) and
elem.is_text_content):
break
output.append(numbered_html[elem.start:elem.end])
output.append(next_text_section)
# Reached the end of the document
except StopIteration:
output.append(numbered_html[element.start:element.end])
else:
for element in target:
output.append(numbered_html[element.start:element.end])
return remove_tagids(''.join(output))
def numbered_html(self):
if hasattr(self, '_numbered_html'):
return self._numbered_html
self._numbered_html = add_tagids(self.html)
return self._numbered_html
def numbered_html(self):
if hasattr(self, '_numbered_html'):
return self._numbered_html
self._numbered_html = add_tagids(self.html)
return self._numbered_html
def port_sample(sample, schemas=None, extractors=None):
"""Convert slybot samples made before slybot 0.13 to new format."""
if schemas is None:
schemas = {}
if extractors is None:
extractors = {}
if sample.get('version') == SLYBOT_VERSION:
return sample, schemas
container_id = gen_predictable_id(sample.get('id', 1), sample['page_id'])
schema_id, schemas = guess_schema(sample, schemas)
default_annotations = [
_create_container('body', container_id, schema_id=schema_id)
]
if not sample.get('annotated_body') and not sample.get('plugins'):
sample['plugins'] = {
'annotations-plugin': {
'extracts': default_annotations
}
}
return sample, schemas
if not sample.get('plugins'):
sample['plugins'] = load_annotations(sample.get('annotated_body', u''))
else:
repair_ids(sample)
# Group annotations by type
annotations = sample['plugins']['annotations-plugin']['extracts']
try:
sel = Selector(text=add_tagids(sample['original_body']))
except KeyError:
annotated = sample.get('annotated_body', u'')
sample['original_body'] = annotated
try:
tagged = add_tagids(annotated)
except KeyError:
tagged = u''
sel = Selector(text=tagged)
sample.pop('annotated_body', None)
annotations = port_standard(annotations, sel, sample, extractors)
standard_annos, generated_annos, variant_annos = [], [], []
for a in annotations:
if a.get('generated'):
generated_annos.append(a)
elif a.get('variants', 0) > 0:
variant_annos.append(a)
else:
standard_annos.append(a)
if not annotations:
sample['plugins'] = {
'annotations-plugin': {
'extracts': default_annotations
}
}
return sample, schemas
new_annotations = []
a = find_element(annotations[0], sel)
for b in annotations[1:]:
b = find_element(b, sel)
a = find_common_parent(a, b)
parent = a.getparent()
container = _create_container(a if parent is None else parent,
container_id,
selector=sel,
schema_id=schema_id)
new_annotations.append(container)
for a in standard_annos:
a.pop('variant', None)
new_annotations.extend(standard_annos)
new_annotations.extend(port_generated(generated_annos, sel))
new_annotations.extend(port_variants(variant_annos, sel))
for a in new_annotations:
if not (a.get('item_container') and a.get('container_id')):
if container_id == a.get('id'):
continue
a['container_id'] = container_id
a.pop('tagid', None) or a.pop('data-tagid', None)
# Update annotations
sample['plugins']['annotations-plugin']['extracts'] = new_annotations
sample['version'] = SLYBOT_VERSION
return sample, schemas
def port_sample(sample, schemas=None, extractors=None):
"""Convert slybot samples made before slybot 0.13 to new format."""
if schemas is None:
schemas = {}
if extractors is None:
extractors = {}
if sample.get('version') == SLYBOT_VERSION:
return sample, schemas
if 'url' not in sample:
sample['url'] = 'http://example.com'
container_id = gen_predictable_id(
sample.get('id', 1), sample.get('page_id', sample['name']))
schema_id, schemas = guess_schema(sample, schemas)
default_annotations = [_create_container('body', container_id,
schema_id=schema_id)]
if not sample.get('annotated_body') and not sample.get('plugins'):
sample['plugins'] = {
'annotations-plugin': {
'extracts': default_annotations
}
}
return sample, schemas
if not sample.get('plugins'):
sample['plugins'] = load_annotations(sample.get('annotated_body', u''))
else:
repair_ids(sample)
# Group annotations by type
annotations = sample['plugins']['annotations-plugin']['extracts']
try:
sel = Selector(text=add_tagids(sample['original_body']))
except KeyError:
annotated = sample.get('annotated_body', u'')
sample['original_body'] = annotated
try:
tagged = add_tagids(annotated)
except KeyError:
tagged = u''
sel = Selector(text=tagged)
sample.pop('annotated_body', None)
annotations = port_standard(annotations, sel, sample, extractors)
standard_annos, generated_annos, variant_annos = [], [], []
for a in annotations:
if a.get('generated'):
generated_annos.append(a)
elif a.get('variants', 0) > 0:
variant_annos.append(a)
else:
standard_annos.append(a)
if not annotations:
sample['plugins'] = {
'annotations-plugin': {
'extracts': default_annotations
}
}
return sample, schemas
new_annotations = []
a = find_element(annotations[0], sel)
for b in annotations[1:]:
b = find_element(b, sel)
a = find_common_parent(a, b)
parent = a.getparent()
container = _create_container(
a if parent is None else parent, container_id, selector=sel,
schema_id=schema_id)
new_annotations.append(container)
for a in standard_annos:
a.pop('variant', None)
new_annotations.extend(standard_annos)
new_annotations.extend(port_generated(generated_annos, sel))
new_annotations.extend(port_variants(variant_annos, sel))
for a in new_annotations:
if not (a.get('item_container') and a.get('container_id')):
if container_id == a.get('id'):
continue
a['container_id'] = container_id
a.pop('tagid', None) or a.pop('data-tagid', None)
# Update annotations
sample['plugins']['annotations-plugin']['extracts'] = new_annotations
sample['version'] = SLYBOT_VERSION
return sample, schemas
