def print_top_ngramms_top_author_by_frags(
mdb, topn:int, *, topn_authors:int=500, nka:int=2, ltype:str='lemmas'
):
"""Кросс-распределение «фразы» - «со-цитирования»"""
print('Кросс-распределение «фразы» - «со-цитирования»')
n_gramms = mdb.n_gramms
top_ngramms = get_topn(
n_gramms, topn, preselect=[{'$match': {'nka': nka, 'type': ltype}}],
sum_expr='$linked_papers.cnt')
contexts = mdb.contexts
for i, (ngramm, cnt, conts) in enumerate(top_ngramms, 1):
frags = Counter()
congr = defaultdict(Counter)
for doc in contexts.aggregate([
{'$match': {'frag_num': {'$gt': 0}, '_id': {'$in': conts}}},
{'$project': {'prefix': False, 'suffix': False, 'exact': False}},
{'$unwind': '$cocit_authors'},
]):
ngr = doc['cocit_authors']
fnum = doc['frag_num']
frags[fnum] += 1
congr[ngr][fnum] += 1
msg = f'{"/".join(str(frags[i]) for i in range(1, 6))}'
print(f"{i:<3d} '{ngramm}' {msg} ({sum(frags.values())})") # cnt})") #
for j, (co, cnts) in enumerate(
sorted(congr.items(), key=lambda kv: (-sum(kv[1].values()), kv[0])), 1):
msg = f'{"/".join(str(cnts[i]) for i in range(1, 6))}'
print(f" {j:<3d} '{co}': {msg} ({sum(cnts.values())})")
def print_topics_top_author_by_frags(mdb):
"""Кросс-распределение «топики» - «со-цитирования»"""
print('Г Кросс-распределение «топики» - «со-цитирования»:',
'topics_top_author_by_frags.json')
top_topics = get_topn(mdb.topics, 100)
contexts = mdb.contexts
out_dict = {}
for i, (topic, cnt, conts) in enumerate(top_topics, 1):
frags = Counter()
congr = defaultdict(Counter)
for doc in contexts.aggregate([
{
'$match': {
'frag_num': {
'$gt': 0
},
'_id': {
'$in': conts
}
}
},
{
'$project': {
'prefix': False,
'suffix': False,
'exact': False
}
},
{
'$unwind': '$cocit_authors'
},
]):
ngr = doc['cocit_authors']
fnum = doc['frag_num']
frags[fnum] += 1
congr[ngr][fnum] += 1
crosscocitaith = {}
out_dict[topic] = dict(sum=cnt,
frags=frags,
cocitaithors=crosscocitaith)
for j, (co, cnts) in enumerate(
sorted(congr.items(),
key=lambda kv: (-sum(kv[1].values()), kv[0])), 1):
crosscocitaith[co] = dict(frags=cnts, sum=sum(cnts.values()))
with open('../out_json/topics_top_author_by_frags.json', 'w') as out:
json.dump(out_dict, out, ensure_ascii=False)
def print_top_author_ngramms_by_frags(
mdb, topn:int, *, topn_gramms:int=500, nka:int=2, ltype:str='lemmas'
):
"""Кросс-распределение «со-цитирования» - «фразы из контекстов цитирований»"""
print('Кросс-распределение «со-цитирования» - «фразы из контекстов цитирований»')
if topn_gramms:
n_gramms = mdb.n_gramms
top_ngramms = get_topn(
n_gramms, topn_gramms, preselect=[{'$match': {'nka': nka, 'type': ltype}}],
sum_expr='$linked_papers.cnt')
exists = frozenset(t for t, _, _ in top_ngramms)
contexts = mdb.contexts
topN = get_topn_cocit_authors(contexts, topn, include_conts=True)
for i, (cocitauthor, cnt, conts) in enumerate(topN, 1):
frags = Counter()
congr = defaultdict(Counter)
for doc in contexts.aggregate([
{'$match': {'frag_num': {'$gt': 0}, '_id': {'$in': conts}}}, # 'cocit_authors': cocitauthor}}, #
{'$project': {'prefix': False, 'suffix': False, 'exact': False}},
{'$lookup': {
'from': 'n_gramms', 'localField': '_id',
'foreignField': 'linked_papers.cont_id', 'as': 'cont'}},
{'$unwind': '$cont'},
{'$match': {'cont.nka': nka, 'cont.type': ltype}},
{'$unwind': '$cont.linked_papers'},
{'$match': {'$expr': {'$eq': ["$_id", "$cont.linked_papers.cont_id"]}}},
{'$project': {'cont.type': False}}, # 'cont.linked_papers': False,
# {'$sort': {'frag_num': 1}},
]):
cont = doc['cont']
ngr = cont['title']
if topn_gramms and ngr not in exists:
continue
fnum = doc['frag_num']
frags[fnum] += 1
congr[ngr][fnum] += cont['linked_papers']['cnt']
msg = f'{"/".join(str(frags[i]) for i in range(1, 6))}'
print(f"{i:<3d} '{cocitauthor}' {msg} ({cnt})") # sum(frags.values())
for j, (co, cnts) in enumerate(
sorted(congr.items(), key=lambda kv: (-sum(kv[1].values()), kv[0])), 1):
msg = f'{"/".join(str(cnts[i]) for i in range(1, 6))}'
print(f" {j:<3d} '{co}': {msg} ({sum(cnts.values())})")
def print_topics_top_author_by_frags(mdb):
"""Кросс-распределение «топики» - «со-цитирования»"""
print('Кросс-распределение «топики» - «со-цитирования»')
top_topics = get_topn(mdb.topics, 100)
contexts = mdb.contexts
for i, (topic, cnt, conts) in enumerate(top_topics, 1):
frags = Counter()
congr = defaultdict(Counter)
for doc in contexts.aggregate([
{
'$match': {
'frag_num': {
'$gt': 0
},
'_id': {
'$in': conts
}
}
},
{
'$project': {
'prefix': False,
'suffix': False,
'exact': False
}
},
{
'$unwind': '$cocit_authors'
},
]):
ngr = doc['cocit_authors']
fnum = doc['frag_num']
frags[fnum] += 1
congr[ngr][fnum] += 1
msg = f'{"/".join(str(frags[i]) for i in range(1, 6))}'
print(f"{i:<3d} '{topic}' {msg} ({sum(frags.values())})") # cnt})") #
for j, (co, cnts) in enumerate(
sorted(congr.items(),
key=lambda kv: (-sum(kv[1].values()), kv[0])), 1):
msg = f'{"/".join(str(cnts[i]) for i in range(1, 6))}'
print(f" {j:<3d} '{co}': {msg} ({sum(cnts.values())})")
def print_top_ngramms_topics_by_frags(
mdb, topn:int=10, *, nka:int=2, ltype:str='lemmas'
):
"""Кросс-распределение «фразы» - «топики контекстов цитирований»"""
print('Кросс-распределение «фразы» - «топики контекстов цитирований»')
n_gramms = mdb.n_gramms
top_ngramms = get_topn(
n_gramms, topn, preselect=[{'$match': {'nka': nka, 'type': ltype}}],
sum_expr='$linked_papers.cnt')
contexts = mdb.contexts
for i, (ngrmm, cnt, conts) in enumerate(top_ngramms, 1):
frags = Counter()
congr = defaultdict(Counter)
for doc in contexts.aggregate([
{'$match': {'frag_num': {'$gt': 0}, '_id': {'$in': conts}}},
{'$project': {'prefix': False, 'suffix': False, 'exact': False}},
{'$lookup': {
'from': 'topics', 'localField': '_id',
'foreignField': 'linked_papers.cont_id', 'as': 'cont'}},
{'$unwind': '$cont'},
{'$unwind': '$cont.linked_papers'},
{'$match': {'$expr': {'$eq': ["$_id", "$cont.linked_papers.cont_id"]}}},
{'$project': {'cont.type': False}}, # 'cont.linked_papers': False,
]):
cont = doc['cont']
ngr = cont['title']
fnum = doc['frag_num']
frags[fnum] += 1
congr[ngr][fnum] += 1
msg = f'{"/".join(str(frags[i]) for i in range(1, 6))}'
print(f"{i:<2d} '{ngrmm}' {msg} ({sum(frags.values())})")
for j, (co, cnts) in enumerate(
sorted(congr.items(), key=lambda kv: (-sum(kv[1].values()), kv[0])), 1):
msg = f'{"/".join(str(cnts[i]) for i in range(1, 6))}'
print(f" {j:<2d} '{co}': {msg} ({sum(cnts.values())})")
def print_top_author_ngramms_by_frags(mdb,
topn: int,
*,
topn_gramms: int = 500,
nka: int = 2,
ltype: str = 'lemmas'):
"""Кросс-распределение «со-цитирования» - «фразы из контекстов цитирований»"""
print(
'Б',
'Кросс-распределение «со-цитирования» - «фразы из контекстов цитирований»:',
'top_author_ngramms_by_frags.json')
if topn_gramms:
n_gramms = mdb.n_gramms
top_ngramms = get_topn(n_gramms,
topn_gramms,
preselect=[{
'$match': {
'nka': nka,
'type': ltype
}
}],
sum_expr='$linked_papers.cnt')
exists = frozenset(t for t, _, _ in top_ngramms)
else:
exists = ()
contexts = mdb.contexts
topN = get_topn_cocit_authors(contexts, topn, include_conts=True)
out_dict = {}
for i, (cocitauthor, cnt, conts) in enumerate(topN, 1):
frags = Counter()
congr = defaultdict(Counter)
for doc in contexts.aggregate([
{
'$match': {
'frag_num': {
'$gt': 0
},
'_id': {
'$in': conts
}
}
}, # 'cocit_authors': cocitauthor}}, #
{
'$project': {
'prefix': False,
'suffix': False,
'exact': False
}
},
{
'$lookup': {
'from': 'n_gramms',
'localField': '_id',
'foreignField': 'linked_papers.cont_id',
'as': 'cont'
}
},
{
'$unwind': '$cont'
},
{
'$match': {
'cont.nka': nka,
'cont.type': ltype
}
},
{
'$unwind': '$cont.linked_papers'
},
{
'$match': {
'$expr': {
'$eq': ["$_id", "$cont.linked_papers.cont_id"]
}
}
},
{
'$project': {
'cont.type': False
}
}, # 'cont.linked_papers': False,
# {'$sort': {'frag_num': 1}},
]):
cont = doc['cont']
ngr = cont['title']
if topn_gramms and ngr not in exists:
continue
fnum = doc['frag_num']
frags[fnum] += 1
congr[ngr][fnum] += cont['linked_papers']['cnt']
crossgrams = {}
out_dict[cocitauthor] = dict(sum=cnt,
frags=frags,
crossgrams=crossgrams)
for j, (co, cnts) in enumerate(
sorted(congr.items(),
key=lambda kv: (-sum(kv[1].values()), kv[0])), 1):
crossgrams[co] = dict(frags=cnts, sum=sum(cnts.values()))
with open('../out_json/top_author_ngramms_by_frags.json', 'w') as out:
json.dump(out_dict, out, ensure_ascii=False)
def print_top_ngramms_topics_by_frags(mdb,
topn: int = 10,
*,
nka: int = 2,
ltype: str = 'lemmas'):
"""Кросс-распределение «фразы» - «топики контекстов цитирований»"""
print('В',
'Кросс-распределение «фразы» - «топики контекстов цитирований»:',
'top_ngramms_topics_by_frags.json')
n_gramms = mdb.n_gramms
top_ngramms = get_topn(n_gramms,
topn,
preselect=[{
'$match': {
'nka': nka,
'type': ltype
}
}],
sum_expr='$linked_papers.cnt')
contexts = mdb.contexts
out_dict = {}
for i, (ngrmm, cnt, conts) in enumerate(top_ngramms, 1):
frags = Counter()
congr = defaultdict(Counter)
for doc in contexts.aggregate([
{
'$match': {
'frag_num': {
'$gt': 0
},
'_id': {
'$in': conts
}
}
},
{
'$project': {
'prefix': False,
'suffix': False,
'exact': False
}
},
{
'$lookup': {
'from': 'topics',
'localField': '_id',
'foreignField': 'linked_papers.cont_id',
'as': 'cont'
}
},
{
'$unwind': '$cont'
},
{
'$unwind': '$cont.linked_papers'
},
{
'$match': {
'$expr': {
'$eq': ["$_id", "$cont.linked_papers.cont_id"]
}
}
},
{
'$project': {
'cont.type': False
}
}, # 'cont.linked_papers': False,
]):
cont = doc['cont']
ngr = cont['title']
fnum = doc['frag_num']
frags[fnum] += 1
congr[ngr][fnum] += 1
crosstopics = {}
out_dict[ngrmm] = dict(sum=cnt, frags=frags, crosstopics=crosstopics)
for j, (co, cnts) in enumerate(
sorted(congr.items(),
key=lambda kv: (-sum(kv[1].values()), kv[0])), 1):
crosstopics[co] = dict(frags=cnts, sum=sum(cnts.values()))
with open('../out_json/top_ngramms_topics_by_frags.json', 'w') as out:
json.dump(out_dict, out, ensure_ascii=False)
def print_top_ngramms_top_author_by_frags(mdb,
topn: int,
*,
topn_authors: int = 500,
nka: int = 2,
ltype: str = 'lemmas'):
"""Кросс-распределение «фразы» - «со-цитирования»"""
print('В', 'Кросс-распределение «фразы» - «со-цитирования»:',
'top_ngramms_top_author_by_frags.json')
n_gramms = mdb.n_gramms
top_ngramms = get_topn(n_gramms,
topn,
preselect=[{
'$match': {
'nka': nka,
'type': ltype
}
}],
sum_expr='$linked_papers.cnt')
contexts = mdb.contexts
out_dict = {}
for i, (ngramm, cnt, conts) in enumerate(top_ngramms, 1):
frags = Counter()
congr = defaultdict(Counter)
for doc in contexts.aggregate([
{
'$match': {
'frag_num': {
'$gt': 0
},
'_id': {
'$in': conts
}
}
},
{
'$project': {
'prefix': False,
'suffix': False,
'exact': False
}
},
{
'$unwind': '$cocit_authors'
},
]):
ngr = doc['cocit_authors']
fnum = doc['frag_num']
frags[fnum] += 1
congr[ngr][fnum] += 1
crosscocitaith = {}
out_dict[ngramm] = dict(sum=cnt,
frags=frags,
cocitaithors=crosscocitaith)
for j, (co, cnts) in enumerate(
sorted(congr.items(),
key=lambda kv: (-sum(kv[1].values()), kv[0])), 1):
crosscocitaith[co] = dict(frags=cnts, sum=sum(cnts.values()))
with open('../out_json/top_ngramms_top_author_by_frags.json', 'w') as out:
json.dump(out_dict, out, ensure_ascii=False)
def print_topics_top_ngramm_by_frags(mdb,
topn_gramms: int = 10,
nka: int = 4,
ltype: str = 'lemmas'):
"""Кросс-распределение «топики» - «фразы»"""
print('Кросс-распределение «топики» - «фразы»')
top_topics = get_topn(mdb.topics, 100)
contexts = mdb.contexts
for i, (topic, cnt, conts) in enumerate(top_topics, 1):
frags = Counter()
congr = defaultdict(Counter)
for doc in contexts.aggregate([
{
'$match': {
'frag_num': {
'$gt': 0
},
'_id': {
'$in': conts
}
}
},
{
'$project': {
'prefix': False,
'suffix': False,
'exact': False
}
},
{
'$lookup': {
'from': 'n_gramms',
'localField': '_id',
'foreignField': 'linked_papers.cont_id',
'as': 'cont'
}
},
{
'$unwind': '$cont'
},
{
'$match': {
'cont.nka': nka,
'cont.type': ltype
}
},
# {'$unwind': '$cont.linked_papers'},
# {'$match': {'$expr': {'$eq': ["$_id", "$cont.linked_papers.cont_id"]}}},
{
'$project': {
'cont.type': False
}
}, # 'cont.linked_papers': False,
{
'$sort': {
'cont.count_in_linked_papers': -1,
'cont.count_all': -1
}
},
# {'$limit': topn_gramms}
]):
cont = doc['cont']
ngr = cont['title']
fnum = doc['frag_num']
frags[fnum] += 1
congr[ngr][fnum] += 1 # cont['linked_papers']['cnt']
if len(congr) == topn_gramms:
break
msg = f'{"/".join(str(frags[i]) for i in range(1, 6))}'
print(f"{i:<3d} '{topic}' {msg} ({sum(frags.values())})")
for j, (co, cnts) in enumerate(
sorted(congr.items(),
key=lambda kv: (-sum(kv[1].values()), kv[0])), 1):
msg = f'{"/".join(str(cnts[i]) for i in range(1, 6))}'
print(f" {j:<3d} '{co}': {msg} ({sum(cnts.values())})")