def assess_doc_cred(doc, cfg):
"""Main credibility assessment for a single doc
:param doc: a validated and normalised document, ready for credibility
assessment
:param cfg: any configs we need to execute/customise the assessment
:returns: a credibility assessment for the doc
:rtype: dict
"""
start = citimings.start()
if content.is_tweet_doc(doc):
result = tweet_credrev.review(doc, cfg)
return result
elif content.is_article_doc(doc):
result = article_credrev.review(doc, cfg)
return result
else:
rev_format = cfg.get('acred_review_format', 'schema.org')
msg = 'Unsupported document (not a %s))' % supported_doc_types
if rev_format == 'cred_assessment':
return {
'@context': ci_context,
'@type': 'DocumentCredibilityAssessment',
'doc_url': doc['url'],
'item_assessed': doc,
'cred_assessment_error': msg,
'date_assessed': isodate.now_utc_timestamp(),
'timings': citimings.timing('assess_doc_cred', start),
'credibility': 0,
'confidence': 0,
'explanation': msg}
else:
rating = {
'@type': 'Rating',
'ratingValue': 0.0,
'confidence': 0.0,
'ratingExplanation': msg}
result = {
'@context': ci_context,
'@type': 'DocumentCredReview',
'reviewAspect': 'credibility',
'itemReviewed': doc,
'dateCreated': isodate.now_utc_timestamp(),
'author': bot_info([], cfg),
'reviewRating': {
**rating,
'identifier': itnorm.calc_identifier(rating, cfg)}
}
return {
**result,
'identifier': itnorm.calc_identifier(result, cfg)
}
def similarSent_as_SentSimilarityReview(simSent, simResult, cfg):
qSent = simResult['q_claim']
simReviewer = simResult['simReviewer']
simVal = simSent['similarity']
return {
'@context':
'http://coinform.eu',
'@type':
'SentSimilarityReview',
'itemReviewed':
content.as_dbq_sentpair(dbSent=simSent['sentence'],
qSent=qSent,
cfg=cfg),
'headline':
simlabel.claim_rel_str(simVal, None),
'reviewRating': {
'@type': 'Rating',
'reviewAspect': 'similarity',
'ratingValue': simVal
},
'dateCreated':
simResult.get('dateCreated', isodate.now_utc_timestamp()),
'author':
simReviewer
}
def calc_domain_credibility(domain, cfg={}):
"""Calculates a `DomainCredibility` for a domain via MisinfoMe
Note that `DomainCredibility` is deprecated, use the `review` method
which produces a `WebSiteCredReview` instead.
:param domain: str e.g. `www.snopes.com`
:returns: a `DomainCredibility`
:rtype: dict
"""
if domain is None:
return default_domain_crediblity(
domain, "Default credibility for unknown domain")
else:
assert type(domain) == str, 'Expecting str, but was %s' (type(domain))
start = citimings.start()
try:
return {
**misinfome_source_credibility(domain), '@context':
'DomainCredibility',
'@type': 'DomainCredibility',
'dateCreated': isodate.now_utc_timestamp(),
'timings': citimings.timing('misinfome_source_credibility',
start)
}
except Exception as e:
logger.error("Failed misinfome source credibility. " + str(e))
return default_domain_crediblity(
domain, "Unable to retrieve credibility assessment")
def review_article(adoc, cfg):
"""Main credibility review for a single article
Refactoring of `assess_article_cred`
:param adoc: analyzed doc as returned by `analyzed_doc`
:param cfg: config to guide this assessment
:returns: a `ArticleCredReview`
:rtype: dict
"""
# TODO: ? start = citimings.start()
domcredReview = adoc_to_website_credReview(adoc, cfg)
content_credReview = review_doc_content_cred(adoc, cfg)
# TODO: ? extract sub_bots from website_cr and aggqsent_cr and make sure it matches default_sub_bots?
agg_rating = aggregate_subReviews(domcredReview, content_credReview, adoc,
cfg)
return {
**base_ArticleCredReview(cfg),
'author':
default_bot_info(cfg),
'dateCreated':
isodate.now_utc_timestamp(),
'itemReviewed':
adoc, # maybe just return ref?
'text':
'%s seems *%s* %s' %
(markdown_ref_for_article(
adoc, cfg), credlabel.rating_label(agg_rating, cfg),
agg_rating.get('ratingExplanation', '(missing explanation)')),
'reviewRating':
agg_rating,
'isBasedOn': [domcredReview, content_credReview]
}
def websiteCredRev_as_qclaimCredRating(websiteCredRev, cfg):
wscr = websiteCredRev
result = {
'@type': 'AggregateRating',
'reviewAspect': 'credibility',
'reviewCount': dictu.get_in(wscr, ['reviewRating', 'reviewCount'], 0),
'ratingCount': dictu.get_in(wscr, ['reviewRating', 'ratingCount'], 0),
'ratingValue': dictu.get_in(wscr, ['reviewRating', 'ratingValue'], 0.0),
'dateCreated': isodate.now_utc_timestamp()
}
if is_by_factchecker(websiteCredRev, cfg):
# reduce domain credibility for fact-checkers, as we want to
# focus on their claim reviews even if their confidence is
# relatively low.
# Refactoring of website_credrev.penalise_credibility
penalty = float(cfg.get('factchecker_website_to_qclaim_confidence_penalty_factor', 0.5))
return {
**result,
'confidence': dictu.get_in(wscr, ['reviewRating', 'confidence'], 0.0) * penalty,
'ratingExplanation': "as it was published in site `%s`. %s %s" % (
dictu.get_in(websiteCredRev, ['itemReviewed', 'name']),
websiteCredRev.get('text', '(Explanation for website credibility missing)'),
"However, the site is a factchecker so it publishes sentences with different credibility values.")
}
else:
return {
**result,
'confidence': dictu.get_in(wscr, ['reviewRating', 'confidence'], 0.0),
'ratingExplanation': "as it was published on site `%s`. %s" % (
dictu.get_in(websiteCredRev, ['itemReviewed', 'name']),
websiteCredRev.get('text', '(Explanation for website credibility missing)'))
}
def base_ArticleCredReview(cfg):
return {
'@context': content.ci_context,
'@type': 'ArticleCredReview',
'additionalType': content.super_types('ArticleCredReview'),
'dateCreated': isodate.now_utc_timestamp(),
}
def search_claim(q_claim):
"""finds similar claims or sentences in a claim database
Finding similar claims or sentences in the co-inform claim database # noqa: E501
:param q_claim: This should be an English sentence or claim. Multiple sentences are not allowed.
:type q_claim: str
:rtype: dict
"""
if type(q_claim) is str:
q_claims = [q_claim]
if type(q_claim) is list:
q_claims = q_claim
if q_claim is None:
raise InvalidUsage("Claim is mandatory")
start = citimings.start()
logger.info('Searching semantic vector space for %s claim(s)' % len(
q_claims))
topn = 5
preds, claim_ids, simReviewer = search_semantic_vecspace(q_claims, topn=topn)
search_semspace_t = citimings.timing('search_semantic_vecspace', start)
assert len(preds) == len(claim_ids)
assert len(q_claims) == len(preds)
q_resp, claim_retrieve_t = retrieve_result_claims(claim_ids, q_claims, topn)
start3 = citimings.start()
results, sub_build_ts = [], []
for i in range(len(q_claims)):
start4 = citimings.start()
claim_id2pred = {idx: float(pred) for idx, pred in zip(
claim_ids[i], preds[i])}
relsents, sub_ts = q_resp_to_related_sent(
q_resp, claim_id2pred)
qclaim = q_claims[i]
results.append({
'@context': ci_context,
'@type': 'SemanticClaimSimilarityResult',
'dateCreated': isodate.now_utc_timestamp(),
'q_claim': qclaim,
'simReviewer': simReviewer,
'results': relsents})
sub_build_ts.append(citimings.timing('build_result', start4, sub_ts))
result_build_t = citimings.timing('build_results', start3, sub_build_ts)
results, stance_pred_t = add_stance_detection(
results, sim_threshold=stance_min_sim_threshold)
timing = citimings.timing(
'search_claim', start,
[search_semspace_t, claim_retrieve_t,
result_build_t, stance_pred_t])
return {
'results': results,
'resultsHeader': {
'QTime': timing['total_ms'],
'timings': timing,
'params': {
'claim': q_claim
}}}
def base_AggQSentCredReview(cfg):
return {
'@context': ci_context,
'@type': 'AggQSentCredReview',
'additionalType': ['CredibilityReview', 'Review'],
'dateCreated': isodate.now_utc_timestamp(),
'author': default_bot_info(cfg) # default sub_bots
}
def claimsim_result_as_claimcred(claimsim_result, cfg):
"""Convert a `SemanticClaimSimilarityResult` into a `ClaimCredibility`
:param claimsim_results:
:param cfg:
:returns:
:rtype:
"""
# TODO: delegate to reviewers to convert claimsim_result into
# QSentCredReview, DBClaimCredibilityReview, WebSiteCredReview, etc.
agg_start = citimings.start()
qsent = claimsim_result['q_claim'] # qsent
relsents = claimsim_result['results'] # simsents
# sentSimReviews = [ # TODO: remove, just for feedback during refactoring
# semsent_simrev.similarSent_as_SentSimilarityReview(simSent, claimsim_result, cfg)
# for simSent in relsents]
for rs in relsents:
# claim search no longer does domain credibility, so we have to do it here
if 'domain_credibility' not in rs:
rs['domain_credibility'] = website_credrev.calc_domain_credibility(
rs['domain'])
relsents = [add_relative_credibility(rs, cfg) for rs in relsents]
cred_dict = aggregate_credibility(relsents, cfg)
cred_dict['source'] = 'credibility of %d related claims ' % len(relsents)
agg_t = citimings.timing('claim_relsent_agg', agg_start)
return {
'@context':
ci_context,
'@type':
'ClaimCredibility',
'claim':
qsent,
'item_assessed': {
'@context': ci_context,
'@type': 'Claim',
'claim': qsent
},
# 'sentenceSimilarityReview': sentSimReviews,
'aggQSentCredReview':
claimsim_result_as_aggQSentCredReview(claimsim_result, cfg),
'related_claims':
_partition_related_sents(relsents, cfg),
'date_assessed':
isodate.now_utc_timestamp(),
'assessor': {
'@context': ci_context,
'@type': 'CredibilityAssessor',
'name': 'SemanticSimilarityClaimCredibilityAssessor',
'version': '20200208'
},
'credibility':
cred_dict,
'timings':
agg_t
}
def normalise(claimReview, cfg):
if claimReview is None:
return None
assert content.is_claimReview(claimReview), "%s" % (claimReview)
sub_ratings = normalised_claimReview_ratings(claimReview)
most_confident = agg.select_most_confident_rating(sub_ratings)
if most_confident is None:
agg_rating = {
'@type':
'AggregateRating',
'reviewAspect':
'credibility',
'reviewCount':
1, # the original claimReview
'ratingCount':
len(sub_ratings),
'ratingValue':
0.0,
'confidence':
0.0,
'ratingExplanation':
'Failed to interpret original [review](claimReview,get("url", "missing_url"))'
}
else:
agg_rating = {
**most_confident, '@type': 'AggregateRating',
'reviewCount': 1,
'ratingCount': len(sub_ratings)
}
assert type(agg_rating['confidence']) == float
assert 'ratingExplanation' in agg_rating, '%s' % (most_confident)
return {
'@context':
ci_context,
'@type':
'NormalisedClaimReview',
'additionalType':
content.super_types('NormalisedClaimReview'),
'author':
bot_info(cfg),
'text':
'Claim `%s` is *%s* %s' %
(claimReview.get('claimReviewed'),
credlabel.rating_label(agg_rating, cfg),
agg_rating.get('ratingExplanation', '(missing explanation)')),
'claimReviewed':
claimReview.get('claimReviewed'),
'dateCreated':
isodate.now_utc_timestamp(),
'isBasedOn': [claimReview] + sub_ratings,
'reviewAspect':
'credibility',
'reviewRating':
agg_rating
}
def from_old_DomainCredibility(dom_cred, cfg):
"""Converts a `DomainCredibility` into a `WebSiteCredReview`
:param dom_cred: a `DomainCredibility` dict
:param cfg: configuration options
:returns: a `WebSiteCredReview`
:rtype: dict
"""
domain_url = dom_cred.get('itemReviewed', 'missing_website') # str
itemReviewed = content.str_as_website(domain_url) # reconstruct WebSite
ratingVal = dictu.get_in(dom_cred, ['credibility', 'value'], 0.0)
explanation = 'based on %d review(s) by external rater(s)%s' % (len(
dom_cred['assessments']), example_raters_markdown(dom_cred))
return {
'@context':
'http://coinform.eu',
'@type':
'WebSiteCredReview',
'additionalType':
content.super_types('WebSiteCredReview'),
'itemReviewed':
itemReviewed,
'text':
'Site `%s` seems *%s* %s' %
(itemReviewed.get('name', '??'),
credlabel.describe_credval(ratingVal, None), explanation),
'author':
misinfoMeSourceCredReviewer(),
'reviewRating': {
'@type':
'AggregateRating',
'reviewAspect':
'credibility',
'ratingValue':
ratingVal,
'confidence':
dictu.get_in(dom_cred, ['credibility', 'confidence'], 0.5),
'ratingExplanation':
explanation,
'reviewCount':
len(dom_cred['assessments']),
'ratingCount':
len(dom_cred['assessments'])
},
'dateCreated':
dom_cred.get('dateCreated', isodate.now_utc_timestamp()),
'reviewAspect':
'credibility',
'isBasedOn': [], # TODO:
'isBasedOn_assessments':
dom_cred['assessments'],
'timings':
dom_cred.get('timings', {})
}
def assess_article_cred(article, cfg):
"""Main credibility assessment for a single article
*Deprecated* you should move to `review_article`
:param article: valid and normalised article
:param cfg: config to guide this assessment
:returns: a credibility assessment for the article
:rtype: dict
"""
start = citimings.start()
adoc = analyzed_doc(article, cfg)
adoc_t = adoc['timings']
domcred = adoc_to_domain_cred(adoc, cfg)
content_cred = assess_doc_content_cred(adoc, cfg)
agg_cred = aggregate_article_cred(domcred, content_cred, cfg)
return {
'@context':
content.ci_context,
'@type':
'ArticleCredibilityAssessment',
'doc_url':
article['url'],
'item_assessed':
article,
'date_asessed':
isodate.now_utc_timestamp(),
'assessor': {
'@context': content.ci_context,
'@type': 'CredibilityAssessor',
'name': 'ArticleCredibilityAssessor',
'version': '20200207'
},
'doc_resolved_url':
adoc.get('resolved_url', adoc.get('url')),
'analyzed_doc':
adoc,
**agg_cred, 'sub_assessments': [domcred, content_cred],
'timings':
citimings.timing('assess_article_cred', start, [
adoc_t,
domcred.get('timings', None),
content_cred.get('timings', None)
])
# 'claims_in_doc': claim_creds,
# 'domain_credibility': domcred,
# 'content_credibility': content_cred
}
def dummyPrediction(tweet):
start = citimings.start()
return {
'@context': ci_context,
'@type': 'TweetCredibilityAssessment',
'tweet_id': int(tweet['tweet_id']),
'item_assessed': tweet,
'credibility': random.random(),
'confidence': 0.0,
'explanation': 'Dummy prediction, no actual analysis performed.',
'sub_assessments': [],
'date_assessed': isodate.now_utc_timestamp(),
'assessor': {'@context': ci_context,
'name': 'dummyCredibilityPredictor'},
'timings': citimings.timing('dummyPrediction', start)
# deprecated, now as sub_assessments
# 'sentences_in_tweets': [],
# 'sentences_linked': []
}
def worthinesspreds_as_SentCheckWorthinessReview(mapped_pred, config):
result = {
"@context": ci_context,
"@type": "SentCheckWorthinessReview",
"additionalType": content.super_types('SentCheckWorthinessReview'),
'reviewAspect': 'checkworthiness',
'itemReviewed': content.as_sentence(mapped_pred['sentence']),
'reviewRating': {
'@type':
'Rating',
'reviewAspect':
'checkworthiness',
'ratingValue':
mapped_pred['ratingValue'],
'confidence':
mapped_pred['confidence'],
'ratingExplanation':
rating_exp(mapped_pred['ratingValue'], mapped_pred['sentence'])
},
'dateCreated': isodate.now_utc_timestamp(),
"author": checkWorthinessReviewer(config)
}
result['identifier'] = calc_worth_review_id(result)
return result
def aggregate_subReviews(db_Sentence, claimReview, webSiteCred, cfg):
"""Aggregates (claim and WebSite) reviews about a DB Sentence into a
credibility review
:param db_Sentence: a `Sentence` in the Co-inform database
:param claimReview: a `ClaimReview` for the db_Sentence. May be
None if no claim review is available for the sentence. In
general, the claim review will not have been normalised
(i.e. mapped onto the co-inform accuracy/credibility scales)
:param webSiteCred: a `WebSiteCredReview` for a webSite where the
`db_Sentence` was published.
:param cfg: configuration options
:returns: a `DBSentCredReview`
:rtype: dict
"""
nClaimReview = crn.normalise(claimReview, cfg)
if nClaimReview is None:
nClaimReview = {}
nWebSiteRating = None
if webSiteCred:
nWebSiteRating = websiteCredRev_as_qclaimCredRating(webSiteCred, cfg)
assert type(nWebSiteRating['confidence']) == float
assert type(dictu.get_in(nClaimReview, ['reviewRating', 'confidence'], 0.0)) == float
subRatings = [nWebSiteRating, nClaimReview.get('reviewRating', None)]
subRatings = [r for r in subRatings if r is not None]
sel_rating = agg.select_most_confident_rating(subRatings) or {
'ratingValue': 0.0,
'confidence': 0.0,
'ratingExplanation': 'No website or claimReview associated with this sentence'
}
isBasedOn = [webSiteCred, nClaimReview]
isBasedOn = [ibo for ibo in isBasedOn
if ibo is not None and ibo != {}]
reviewCount = agg.total_reviewCount(subRatings) + len(isBasedOn)
ratingCount = agg.total_ratingCount(subRatings)
# should be a superset of [ibo.get('author') for ibo in isBasedOn]
sub_bots = default_sub_bots(cfg)
appears_in_docs = db_Sentence.get('appearance', [])
appears_in_doc = appears_in_docs[0] if appears_in_docs else None
link_to_doc = md_link_to_doc(appears_in_doc)
revRating = {
'@type': 'AggregateRating',
'reviewAspect': 'credibility',
'reviewCount': reviewCount,
'ratingCount': ratingCount,
'ratingValue': sel_rating.get('ratingValue', 0.0),
'confidence': sel_rating.get('confidence', 0.0),
'ratingExplanation': sel_rating.get('ratingExplanation')
}
return {
'@context': "http://coinform.eu",
'@type': "DBSentCredReview",
'additionalType': content.super_types('DBSentCredReview'),
'itemReviewed': db_Sentence,
'text': 'Sentence `%s` %sseems *%s* %s' % (
db_Sentence.get('text', '??'),
', in %s, ' % (link_to_doc) if link_to_doc else '',
credlabel.rating_label(revRating, cfg),
sel_rating.get('ratingExplanation')
),
'reviewRating': revRating,
'reviewAspect': 'credibility',
'isBasedOn': isBasedOn,
'dateCreated': isodate.now_utc_timestamp(),
'author': bot_info(sub_bots, cfg)
}
def aggregate_subReviews(subReviews, tweet, cfg):
"""Creates an aggregate review based on subReviews for tweet
Refactoring of `aggregate_tweet_cred`
:param subReviews: list of credibility reviews for (parts of) the
tweet to review.
:param cfg: config options
:returns: a credibility review for the `tweet` to review that
contains an `AggregateRating` based on the `subReviews`
:rtype: dict
"""
# extract sub_bots and compare to default_sub_bots
partial_TweetCredReview = {
'@context': ci_context,
'@type': 'TweetCredReview',
'itemReviewed': tweet,
'isBasedOn': subReviews,
'dateCreated': isodate.now_utc_timestamp(),
'author': default_bot_info(cfg)
}
tweet_mdref = markdown_ref_for_tweet(tweet, cfg)
if subReviews is None:
subReviews = []
subRatings = [
sr.get('reviewRating') for sr in subReviews
if sr.get('reviewRating') is not None
]
# filter by min confidence
conf_threshold = float(cfg.get('cred_conf_threshold', 0.7))
filter_fn = agg.filter_review_by_min_confidence(conf_threshold)
conf_subRevs = [sr for sr in subReviews if filter_fn(sr)]
igno_subRevs = [sr for sr in subReviews if not filter_fn(sr)]
# no (confident) subReviews
if len(conf_subRevs) == 0:
part_rating = {
'@type': 'Rating',
'ratingValue': 0.0,
'confidence': 0.0,
'reviewAspect': 'credibility'
}
if len(subReviews) == 0:
msg = "we could not extract (or assess credibility of) its sentences or linked documents" % (
tweet_mdref)
rating = {**part_rating, 'ratingExplanation': msg}
else:
msg = 'we could not assess the credibility of its %d sentences or linked documents.%s' % (
len(subReviews), '\nFor example:\n * %s' %
(igno_subRevs[0]['text']))
rating = {
**part_rating, '@type': 'AggregateRating',
'ratingExplanation': msg,
'ratingCount': agg.total_ratingCount(subRatings),
'reviewCount':
agg.total_reviewCount(subRatings) + len(subReviews)
}
return {
**partial_TweetCredReview, 'text':
'%s seems *%s* as %s' %
(tweet_mdref, credlabel.rating_label(rating, cfg), msg),
'reviewRating':
rating
}
# select least credible subReview
subRevs_by_val = sorted([sr for sr in conf_subRevs],
key=lambda rev: dictu.get_in(
rev, ['reviewRating', 'ratingValue'], 0.0))
least_cred_rev = subRevs_by_val[0]
msg = 'based on its least credible part:\n%s' % (dictu.get_in(
least_cred_rev, ['text'], '(missing explanation for part)'))
revRating = {
'@type':
'AggregateRating',
'reviewAspect':
'credibility',
'ratingValue':
dictu.get_in(least_cred_rev, ['reviewRating', 'ratingValue'], 0.0),
'confidence':
dictu.get_in(least_cred_rev, ['reviewRating', 'confidence'], 0.0),
'ratingExplanation':
msg,
'ratingCount':
agg.total_ratingCount(subRatings),
'reviewCount':
agg.total_reviewCount(subRatings) + len(subReviews)
}
return {
**partial_TweetCredReview,
'isBasedOn':
subRevs_by_val + igno_subRevs, # just a re-ordering
'text':
'%s seems *%s* %s' %
(tweet_mdref, credlabel.rating_label(revRating, cfg), msg),
'reviewRating':
revRating
}
def aggregate_subReviews(simple_sentSimReview, stanceReview, cfg):
"""Aggregates a similarity and stance review into a polar similarity review
:param simple_sentSimReview: a (non-polar) `SentSimilarityReview`
for a `sentPair`
:param stanceReview: a `SentStanceReview` for the same `sentPair`
as `simple_sentSimReview`
:param cfg: configuration options
:returns: a `SentPolarSimilarityReview`
:rtype: dict
"""
assert simple_sentSimReview is not None
if stanceReview is None:
return simple_sentSimReview
sim = dictu.get_in(simple_sentSimReview, ['reviewRating', 'ratingValue'])
sent_stance = dictu.get_in(stanceReview, ['reviewRating', 'ratingValue'],
'unrelated')
stance_conf = dictu.get_in(stanceReview, ['reviewRating', 'confidence'],
'0.5')
sent_pair = simple_sentSimReview['itemReviewed']
assert stanceReview['itemReviewed'] == sent_pair, '%s != %s' % (
stanceReview['itemReviewed'], sent_pair)
agg_sim = calc_agg_polarsim(sim=sim,
sent_stance=sent_stance,
sent_stance_conf=stance_conf,
cfg=cfg)
sub_reviews = [
sr for sr in [simple_sentSimReview, stanceReview] if sr is not None
]
sub_ratings = [
srev.get('reviewRating') for srev in sub_reviews
if srev.get('reviewRating') is not None
]
headline = simlabel.claim_rel_str(sim, sent_stance)
# TODO: more than an explanation this is the review body
# the explanation would be that one model said the sentences were x similar
# while another said they were (stance)
explanation = 'Sentence `%s` %s `%s`' % (dictu.get_in(
sent_pair, ['sentA', 'text'
]), headline, dictu.get_in(sent_pair, ['sentB', 'text']))
sub_bots = [
simple_sentSimReview.get('author', {}),
stanceReview.get('author', {})
]
return {
'@context': 'http://coinform.eu',
'@type': 'SentPolarSimilarityReview',
'additionalType': content.super_types('SentPolarSimilarityReview'),
'itemReviewed': sent_pair,
'headline': headline,
'reviewAspect': 'polarSimilarity',
'reviewBody': explanation,
'reviewRating': {
'@type': 'AggregateRating',
'reviewAspect': 'polarSimilarity',
'ratingValue': agg_sim,
'confidence': stance_conf,
'reviewCount': len(sub_reviews),
'ratingCount': agg.total_ratingCount(sub_ratings),
'ratingExplanation': explanation
},
'isBasedOn': sub_reviews,
'dateCreated': isodate.now_utc_timestamp(),
'author': bot_info(sub_bots, cfg)
}