Spaces:
Build error
Build error
merging dataset statistics file
Browse files- data_measurements/dataset_statistics.py +112 -217
data_measurements/dataset_statistics.py
CHANGED
@@ -15,11 +15,12 @@
|
|
15 |
import json
|
16 |
import logging
|
17 |
import statistics
|
18 |
-
import torch
|
19 |
from os import mkdir
|
20 |
from os.path import exists, isdir
|
21 |
from os.path import join as pjoin
|
22 |
|
|
|
|
|
23 |
import nltk
|
24 |
import numpy as np
|
25 |
import pandas as pd
|
@@ -28,31 +29,17 @@ import plotly.express as px
|
|
28 |
import plotly.figure_factory as ff
|
29 |
import plotly.graph_objects as go
|
30 |
import pyarrow.feather as feather
|
31 |
-
import matplotlib.pyplot as plt
|
32 |
-
import matplotlib.image as mpimg
|
33 |
import seaborn as sns
|
|
|
34 |
from datasets import load_from_disk
|
35 |
from nltk.corpus import stopwords
|
36 |
from sklearn.feature_extraction.text import CountVectorizer
|
37 |
|
38 |
-
from .dataset_utils import (
|
39 |
-
|
40 |
-
TOT_OPEN_WORDS,
|
41 |
-
|
42 |
-
|
43 |
-
EMBEDDING_FIELD,
|
44 |
-
LENGTH_FIELD,
|
45 |
-
OUR_LABEL_FIELD,
|
46 |
-
OUR_TEXT_FIELD,
|
47 |
-
PROP,
|
48 |
-
TEXT_NAN_CNT,
|
49 |
-
TOKENIZED_FIELD,
|
50 |
-
TXT_LEN,
|
51 |
-
VOCAB,
|
52 |
-
WORD,
|
53 |
-
extract_field,
|
54 |
-
load_truncated_dataset,
|
55 |
-
)
|
56 |
from .embeddings import Embeddings
|
57 |
from .npmi import nPMI
|
58 |
from .zipf import Zipf
|
@@ -151,6 +138,7 @@ _NUM_VOCAB_BATCHES = 2000
|
|
151 |
_TOP_N = 100
|
152 |
_CVEC = CountVectorizer(token_pattern="(?u)\\b\\w+\\b", lowercase=True)
|
153 |
|
|
|
154 |
class DatasetStatisticsCacheClass:
|
155 |
def __init__(
|
156 |
self,
|
@@ -249,13 +237,13 @@ class DatasetStatisticsCacheClass:
|
|
249 |
# path to the directory used for caching
|
250 |
if not isinstance(text_field, str):
|
251 |
text_field = "-".join(text_field)
|
252 |
-
#if isinstance(label_field, str):
|
253 |
# label_field = label_field
|
254 |
-
#else:
|
255 |
# label_field = "-".join(label_field)
|
256 |
self.cache_path = pjoin(
|
257 |
self.cache_dir,
|
258 |
-
f"{dset_name}_{dset_config}_{split_name}_{text_field}",
|
259 |
)
|
260 |
if not isdir(self.cache_path):
|
261 |
logs.warning("Creating cache directory %s." % self.cache_path)
|
@@ -284,14 +272,15 @@ class DatasetStatisticsCacheClass:
|
|
284 |
# Needed for UI
|
285 |
self.dup_counts_df_fid = pjoin(self.cache_path, "dup_counts_df.feather")
|
286 |
# Needed for UI
|
287 |
-
self.fig_tok_length_fid = pjoin(self.cache_path, "fig_tok_length.
|
288 |
|
289 |
## General text stats
|
290 |
# Needed for UI
|
291 |
self.general_stats_json_fid = pjoin(self.cache_path, "general_stats_dict.json")
|
292 |
# Needed for UI
|
293 |
-
self.sorted_top_vocab_df_fid = pjoin(
|
294 |
-
|
|
|
295 |
## Zipf cache files
|
296 |
# Needed for UI
|
297 |
self.zipf_fid = pjoin(self.cache_path, "zipf_basic_stats.json")
|
@@ -303,7 +292,6 @@ class DatasetStatisticsCacheClass:
|
|
303 |
self.node_list_fid = pjoin(self.cache_path, "node_list.th")
|
304 |
# Needed for UI
|
305 |
self.fig_tree_json_fid = pjoin(self.cache_path, "fig_tree.json")
|
306 |
-
self.zipf_counts = None
|
307 |
|
308 |
self.live = False
|
309 |
|
@@ -343,18 +331,17 @@ class DatasetStatisticsCacheClass:
|
|
343 |
and exists(self.dup_counts_df_fid)
|
344 |
and exists(self.sorted_top_vocab_df_fid)
|
345 |
):
|
346 |
-
logs.info(
|
347 |
self.load_general_stats()
|
348 |
else:
|
349 |
if not self.live:
|
350 |
-
logs.info(
|
351 |
self.prepare_general_stats()
|
352 |
if save:
|
353 |
write_df(self.sorted_top_vocab_df, self.sorted_top_vocab_df_fid)
|
354 |
write_df(self.dup_counts_df, self.dup_counts_df_fid)
|
355 |
write_json(self.general_stats_dict, self.general_stats_json_fid)
|
356 |
|
357 |
-
|
358 |
def load_or_prepare_text_lengths(self, save=True):
|
359 |
"""
|
360 |
The text length widget relies on this function, which provides
|
@@ -366,15 +353,13 @@ class DatasetStatisticsCacheClass:
|
|
366 |
|
367 |
"""
|
368 |
# Text length figure
|
369 |
-
if
|
370 |
self.fig_tok_length_png = mpimg.imread(self.fig_tok_length_fid)
|
371 |
-
self.fig_tok_length = read_plotly(self.fig_tok_length_fid)
|
372 |
else:
|
373 |
if not self.live:
|
374 |
self.prepare_fig_text_lengths()
|
375 |
if save:
|
376 |
-
|
377 |
-
|
378 |
# Text length dataframe
|
379 |
if self.use_cache and exists(self.length_df_fid):
|
380 |
self.length_df = feather.read_feather(self.length_df_fid)
|
@@ -401,51 +386,48 @@ class DatasetStatisticsCacheClass:
|
|
401 |
if not self.live:
|
402 |
if self.tokenized_df is None:
|
403 |
self.tokenized_df = self.do_tokenization()
|
404 |
-
self.tokenized_df[LENGTH_FIELD] = self.tokenized_df[
|
405 |
-
|
406 |
-
self.length_df = self.tokenized_df[
|
407 |
-
[LENGTH_FIELD, OUR_TEXT_FIELD]].sort_values(
|
408 |
-
by=[LENGTH_FIELD], ascending=True
|
409 |
)
|
|
|
|
|
|
|
410 |
|
411 |
def prepare_text_length_stats(self):
|
412 |
if not self.live:
|
413 |
-
if
|
|
|
|
|
|
|
|
|
414 |
self.prepare_length_df()
|
415 |
-
avg_length = sum(self.tokenized_df[LENGTH_FIELD])/len(
|
|
|
|
|
416 |
self.avg_length = round(avg_length, 1)
|
417 |
std_length = statistics.stdev(self.tokenized_df[LENGTH_FIELD])
|
418 |
self.std_length = round(std_length, 1)
|
419 |
self.num_uniq_lengths = len(self.length_df["length"].unique())
|
420 |
-
self.length_stats_dict = {
|
421 |
-
|
422 |
-
|
|
|
|
|
423 |
|
424 |
def prepare_fig_text_lengths(self):
|
425 |
if not self.live:
|
426 |
-
if
|
|
|
|
|
|
|
427 |
self.prepare_length_df()
|
428 |
self.fig_tok_length = make_fig_lengths(self.tokenized_df, LENGTH_FIELD)
|
429 |
|
430 |
-
def load_or_prepare_embeddings(self
|
431 |
-
|
432 |
-
|
433 |
-
|
434 |
-
|
435 |
-
self.node_list = torch.load(self.node_list_fid)
|
436 |
-
self.fig_tree = make_tree_plot(self.node_list,
|
437 |
-
self.text_dset)
|
438 |
-
if save:
|
439 |
-
write_plotly(self.fig_tree, self.fig_tree_json_fid)
|
440 |
-
else:
|
441 |
-
self.embeddings = Embeddings(self, use_cache=self.use_cache)
|
442 |
-
self.embeddings.make_hierarchical_clustering()
|
443 |
-
self.node_list = self.embeddings.node_list
|
444 |
-
self.fig_tree = make_tree_plot(self.node_list,
|
445 |
-
self.text_dset)
|
446 |
-
if save:
|
447 |
-
torch.save(self.node_list, self.node_list_fid)
|
448 |
-
write_plotly(self.fig_tree, self.fig_tree_json_fid)
|
449 |
|
450 |
# get vocab with word counts
|
451 |
def load_or_prepare_vocab(self, save=True):
|
@@ -455,10 +437,7 @@ class DatasetStatisticsCacheClass:
|
|
455 |
:param
|
456 |
:return:
|
457 |
"""
|
458 |
-
if (
|
459 |
-
self.use_cache
|
460 |
-
and exists(self.vocab_counts_df_fid)
|
461 |
-
):
|
462 |
logs.info("Reading vocab from cache")
|
463 |
self.load_vocab()
|
464 |
self.vocab_counts_filtered_df = filter_vocab(self.vocab_counts_df)
|
@@ -505,7 +484,9 @@ class DatasetStatisticsCacheClass:
|
|
505 |
write_df(self.dup_counts_df, self.dup_counts_df_fid)
|
506 |
|
507 |
def load_general_stats(self):
|
508 |
-
self.general_stats_dict = json.load(
|
|
|
|
|
509 |
with open(self.sorted_top_vocab_df_fid, "rb") as f:
|
510 |
self.sorted_top_vocab_df = feather.read_feather(f)
|
511 |
self.text_nan_count = self.general_stats_dict[TEXT_NAN_CNT]
|
@@ -540,8 +521,7 @@ class DatasetStatisticsCacheClass:
|
|
540 |
if not self.live:
|
541 |
if self.tokenized_df is None:
|
542 |
self.load_or_prepare_tokenized_df()
|
543 |
-
dup_df = self.tokenized_df[
|
544 |
-
self.tokenized_df.duplicated([OUR_TEXT_FIELD])]
|
545 |
self.dup_counts_df = pd.DataFrame(
|
546 |
dup_df.pivot_table(
|
547 |
columns=[OUR_TEXT_FIELD], aggfunc="size"
|
@@ -581,7 +561,7 @@ class DatasetStatisticsCacheClass:
|
|
581 |
write_json({"dset peek": self.dset_peek}, self.dset_peek_json_fid)
|
582 |
|
583 |
def load_or_prepare_tokenized_df(self, save=True):
|
584 |
-
if
|
585 |
self.tokenized_df = feather.read_feather(self.tokenized_df_fid)
|
586 |
else:
|
587 |
if not self.live:
|
@@ -593,7 +573,7 @@ class DatasetStatisticsCacheClass:
|
|
593 |
write_df(self.tokenized_df, self.tokenized_df_fid)
|
594 |
|
595 |
def load_or_prepare_text_dset(self, save=True):
|
596 |
-
if
|
597 |
# load extracted text
|
598 |
self.text_dset = load_from_disk(self.text_dset_fid)
|
599 |
logs.warning("Loaded dataset from disk")
|
@@ -711,8 +691,6 @@ class DatasetStatisticsCacheClass:
|
|
711 |
zipf_dict = json.load(f)
|
712 |
self.z = Zipf()
|
713 |
self.z.load(zipf_dict)
|
714 |
-
# TODO: Should this be cached?
|
715 |
-
self.zipf_counts = self.z.calc_zipf_counts(self.vocab_counts_df)
|
716 |
self.zipf_fig = read_plotly(self.zipf_fig_fid)
|
717 |
elif self.use_cache and exists(self.zipf_fid):
|
718 |
# TODO: Read zipf data so that the vocab is there.
|
@@ -775,30 +753,26 @@ class nPMIStatisticsCacheClass:
|
|
775 |
and exists(self.npmi_terms_fid)
|
776 |
and json.load(open(self.npmi_terms_fid))["available terms"] != []
|
777 |
):
|
778 |
-
|
779 |
else:
|
780 |
-
|
781 |
-
|
782 |
-
|
783 |
-
|
784 |
-
|
785 |
-
|
786 |
-
|
787 |
-
|
788 |
-
|
789 |
-
|
790 |
-
|
791 |
-
|
792 |
-
|
793 |
-
|
794 |
-
|
795 |
-
|
796 |
-
|
797 |
-
|
798 |
-
self.available_terms = available_terms
|
799 |
-
return self.available_terms
|
800 |
-
|
801 |
-
def load_or_prepare_joint_npmi(self, subgroup_pair, save=True):
|
802 |
"""
|
803 |
Run on-the fly, while the app is already open,
|
804 |
as it depends on the subgroup terms that the user chooses
|
@@ -823,7 +797,13 @@ class nPMIStatisticsCacheClass:
|
|
823 |
# When everything is already computed for the selected subgroups.
|
824 |
logs.info("Loading cached joint npmi")
|
825 |
joint_npmi_df = self.load_joint_npmi_df(joint_npmi_fid)
|
826 |
-
npmi_display_cols = [
|
|
|
|
|
|
|
|
|
|
|
|
|
827 |
joint_npmi_df = joint_npmi_df[npmi_display_cols]
|
828 |
# When maybe some things have been computed for the selected subgroups.
|
829 |
else:
|
@@ -832,14 +812,12 @@ class nPMIStatisticsCacheClass:
|
|
832 |
joint_npmi_df, subgroup_dict = self.prepare_joint_npmi_df(
|
833 |
subgroup_pair, subgroup_files
|
834 |
)
|
835 |
-
|
836 |
-
|
837 |
-
|
838 |
-
|
839 |
-
|
840 |
-
|
841 |
-
with open(joint_npmi_fid, "w+") as f:
|
842 |
-
joint_npmi_df.to_csv(f)
|
843 |
else:
|
844 |
joint_npmi_df = pd.DataFrame()
|
845 |
logs.info("The joint npmi df is")
|
@@ -881,7 +859,7 @@ class nPMIStatisticsCacheClass:
|
|
881 |
subgroup_dict[subgroup] = cached_results
|
882 |
logs.info("Calculating for subgroup list")
|
883 |
joint_npmi_df, subgroup_dict = self.do_npmi(subgroup_pair, subgroup_dict)
|
884 |
-
return joint_npmi_df, subgroup_dict
|
885 |
|
886 |
# TODO: Update pairwise assumption
|
887 |
def do_npmi(self, subgroup_pair, subgroup_dict):
|
@@ -892,7 +870,6 @@ class nPMIStatisticsCacheClass:
|
|
892 |
:return: Selected identity term's co-occurrence counts with
|
893 |
other words, pmi per word, and nPMI per word.
|
894 |
"""
|
895 |
-
no_results = False
|
896 |
logs.info("Initializing npmi class")
|
897 |
npmi_obj = self.set_npmi_obj()
|
898 |
# Canonical ordering used
|
@@ -900,26 +877,18 @@ class nPMIStatisticsCacheClass:
|
|
900 |
# Calculating nPMI statistics
|
901 |
for subgroup in subgroup_pair:
|
902 |
# If the subgroup data is already computed, grab it.
|
903 |
-
# TODO: Should we set idx and column names similarly to
|
904 |
-
# how we set them for cached files?
|
905 |
if subgroup not in subgroup_dict:
|
906 |
logs.info("Calculating statistics for %s" % subgroup)
|
907 |
vocab_cooc_df, pmi_df, npmi_df = npmi_obj.calc_metrics(subgroup)
|
908 |
-
|
909 |
-
|
910 |
-
|
911 |
-
|
912 |
-
|
913 |
-
|
914 |
-
|
915 |
-
|
916 |
-
else:
|
917 |
-
# Pair the subgroups together, indexed by all words that
|
918 |
-
# co-occur between them.
|
919 |
-
logs.info("Computing pairwise npmi bias")
|
920 |
-
paired_results = npmi_obj.calc_paired_metrics(subgroup_pair, subgroup_dict)
|
921 |
-
UI_results = make_npmi_fig(paired_results, subgroup_pair)
|
922 |
-
return UI_results.dropna(), subgroup_dict
|
923 |
|
924 |
def set_npmi_obj(self):
|
925 |
"""
|
@@ -993,9 +962,11 @@ class nPMIStatisticsCacheClass:
|
|
993 |
def get_available_terms(self):
|
994 |
return self.load_or_prepare_npmi_terms()
|
995 |
|
|
|
996 |
def dummy(doc):
|
997 |
return doc
|
998 |
|
|
|
999 |
def count_vocab_frequencies(tokenized_df):
|
1000 |
"""
|
1001 |
Based on an input pandas DataFrame with a 'text' column,
|
@@ -1010,7 +981,9 @@ def count_vocab_frequencies(tokenized_df):
|
|
1010 |
)
|
1011 |
# We do this to calculate per-word statistics
|
1012 |
# Fast calculation of single word counts
|
1013 |
-
logs.info(
|
|
|
|
|
1014 |
cvec.fit(tokenized_df[TOKENIZED_FIELD])
|
1015 |
document_matrix = cvec.transform(tokenized_df[TOKENIZED_FIELD])
|
1016 |
batches = np.linspace(0, tokenized_df.shape[0], _NUM_VOCAB_BATCHES).astype(int)
|
@@ -1031,6 +1004,7 @@ def count_vocab_frequencies(tokenized_df):
|
|
1031 |
word_count_df.index.name = WORD
|
1032 |
return word_count_df
|
1033 |
|
|
|
1034 |
def calc_p_word(word_count_df):
|
1035 |
# p(word)
|
1036 |
word_count_df[PROP] = word_count_df[CNT] / float(sum(word_count_df[CNT]))
|
@@ -1041,8 +1015,7 @@ def calc_p_word(word_count_df):
|
|
1041 |
|
1042 |
def filter_vocab(vocab_counts_df):
|
1043 |
# TODO: Add warnings (which words are missing) to log file?
|
1044 |
-
filtered_vocab_counts_df = vocab_counts_df.drop(_CLOSED_CLASS,
|
1045 |
-
errors="ignore")
|
1046 |
filtered_count = filtered_vocab_counts_df[CNT]
|
1047 |
filtered_count_denom = float(sum(filtered_vocab_counts_df[CNT]))
|
1048 |
filtered_vocab_counts_df[PROP] = filtered_count / filtered_count_denom
|
@@ -1051,19 +1024,23 @@ def filter_vocab(vocab_counts_df):
|
|
1051 |
|
1052 |
## Figures ##
|
1053 |
|
|
|
1054 |
def write_plotly(fig, fid):
|
1055 |
write_json(plotly.io.to_json(fig), fid)
|
1056 |
|
|
|
1057 |
def read_plotly(fid):
|
1058 |
fig = plotly.io.from_json(json.load(open(fid, encoding="utf-8")))
|
1059 |
return fig
|
1060 |
|
|
|
1061 |
def make_fig_lengths(tokenized_df, length_field):
|
1062 |
-
fig_tok_length =
|
1063 |
-
|
1064 |
-
)
|
1065 |
return fig_tok_length
|
1066 |
|
|
|
1067 |
def make_fig_labels(label_df, label_names, label_field):
|
1068 |
labels = label_df[label_field].unique()
|
1069 |
label_sums = [len(label_df[label_df[label_field] == label]) for label in labels]
|
@@ -1144,89 +1121,6 @@ def make_zipf_fig(vocab_counts_df, z):
|
|
1144 |
return fig
|
1145 |
|
1146 |
|
1147 |
-
def make_tree_plot(node_list, text_dset):
|
1148 |
-
nid_map = dict([(node["nid"], nid) for nid, node in enumerate(node_list)])
|
1149 |
-
|
1150 |
-
for nid, node in enumerate(node_list):
|
1151 |
-
node["label"] = node.get(
|
1152 |
-
"label",
|
1153 |
-
f"{nid:2d} - {node['weight']:5d} items <br>"
|
1154 |
-
+ "<br>".join(
|
1155 |
-
[
|
1156 |
-
"> " + txt[:64] + ("..." if len(txt) >= 63 else "")
|
1157 |
-
for txt in list(
|
1158 |
-
set(text_dset.select(node["example_ids"])[OUR_TEXT_FIELD])
|
1159 |
-
)[:5]
|
1160 |
-
]
|
1161 |
-
),
|
1162 |
-
)
|
1163 |
-
|
1164 |
-
# make plot nodes
|
1165 |
-
# TODO: something more efficient than set to remove duplicates
|
1166 |
-
labels = [node["label"] for node in node_list]
|
1167 |
-
|
1168 |
-
root = node_list[0]
|
1169 |
-
root["X"] = 0
|
1170 |
-
root["Y"] = 0
|
1171 |
-
|
1172 |
-
def rec_make_coordinates(node):
|
1173 |
-
total_weight = 0
|
1174 |
-
add_weight = len(node["example_ids"]) - sum(
|
1175 |
-
[child["weight"] for child in node["children"]]
|
1176 |
-
)
|
1177 |
-
for child in node["children"]:
|
1178 |
-
child["X"] = node["X"] + total_weight
|
1179 |
-
child["Y"] = node["Y"] - 1
|
1180 |
-
total_weight += child["weight"] + add_weight / len(node["children"])
|
1181 |
-
rec_make_coordinates(child)
|
1182 |
-
|
1183 |
-
rec_make_coordinates(root)
|
1184 |
-
|
1185 |
-
E = [] # list of edges
|
1186 |
-
Xn = []
|
1187 |
-
Yn = []
|
1188 |
-
Xe = []
|
1189 |
-
Ye = []
|
1190 |
-
for nid, node in enumerate(node_list):
|
1191 |
-
Xn += [node["X"]]
|
1192 |
-
Yn += [node["Y"]]
|
1193 |
-
for child in node["children"]:
|
1194 |
-
E += [(nid, nid_map[child["nid"]])]
|
1195 |
-
Xe += [node["X"], child["X"], None]
|
1196 |
-
Ye += [node["Y"], child["Y"], None]
|
1197 |
-
|
1198 |
-
# make figure
|
1199 |
-
fig = go.Figure()
|
1200 |
-
fig.add_trace(
|
1201 |
-
go.Scatter(
|
1202 |
-
x=Xe,
|
1203 |
-
y=Ye,
|
1204 |
-
mode="lines",
|
1205 |
-
line=dict(color="rgb(210,210,210)", width=1),
|
1206 |
-
hoverinfo="none",
|
1207 |
-
)
|
1208 |
-
)
|
1209 |
-
fig.add_trace(
|
1210 |
-
go.Scatter(
|
1211 |
-
x=Xn,
|
1212 |
-
y=Yn,
|
1213 |
-
mode="markers",
|
1214 |
-
name="nodes",
|
1215 |
-
marker=dict(
|
1216 |
-
symbol="circle-dot",
|
1217 |
-
size=18,
|
1218 |
-
color="#6175c1",
|
1219 |
-
line=dict(color="rgb(50,50,50)", width=1)
|
1220 |
-
# '#DB4551',
|
1221 |
-
),
|
1222 |
-
text=labels,
|
1223 |
-
hoverinfo="text",
|
1224 |
-
opacity=0.8,
|
1225 |
-
)
|
1226 |
-
)
|
1227 |
-
return fig
|
1228 |
-
|
1229 |
-
|
1230 |
## Input/Output ###
|
1231 |
|
1232 |
|
@@ -1280,6 +1174,7 @@ def write_json(json_dict, json_fid):
|
|
1280 |
with open(json_fid, "w", encoding="utf-8") as f:
|
1281 |
json.dump(json_dict, f)
|
1282 |
|
|
|
1283 |
def write_subgroup_npmi_data(subgroup, subgroup_dict, subgroup_files):
|
1284 |
"""
|
1285 |
Saves the calculated nPMI statistics to their output files.
|
@@ -1299,6 +1194,7 @@ def write_subgroup_npmi_data(subgroup, subgroup_dict, subgroup_files):
|
|
1299 |
with open(subgroup_cooc_fid, "w+") as f:
|
1300 |
subgroup_cooc_df.to_csv(f)
|
1301 |
|
|
|
1302 |
def write_zipf_data(z, zipf_fid):
|
1303 |
zipf_dict = {}
|
1304 |
zipf_dict["xmin"] = int(z.xmin)
|
@@ -1310,4 +1206,3 @@ def write_zipf_data(z, zipf_fid):
|
|
1310 |
zipf_dict["uniq_ranks"] = [int(rank) for rank in z.uniq_ranks]
|
1311 |
with open(zipf_fid, "w+", encoding="utf-8") as f:
|
1312 |
json.dump(zipf_dict, f)
|
1313 |
-
|
|
|
15 |
import json
|
16 |
import logging
|
17 |
import statistics
|
|
|
18 |
from os import mkdir
|
19 |
from os.path import exists, isdir
|
20 |
from os.path import join as pjoin
|
21 |
|
22 |
+
import matplotlib.pyplot as plt
|
23 |
+
import matplotlib.image as mpimg
|
24 |
import nltk
|
25 |
import numpy as np
|
26 |
import pandas as pd
|
|
|
29 |
import plotly.figure_factory as ff
|
30 |
import plotly.graph_objects as go
|
31 |
import pyarrow.feather as feather
|
|
|
|
|
32 |
import seaborn as sns
|
33 |
+
import torch
|
34 |
from datasets import load_from_disk
|
35 |
from nltk.corpus import stopwords
|
36 |
from sklearn.feature_extraction.text import CountVectorizer
|
37 |
|
38 |
+
from .dataset_utils import (CNT, DEDUP_TOT, EMBEDDING_FIELD, LENGTH_FIELD,
|
39 |
+
OUR_LABEL_FIELD, OUR_TEXT_FIELD, PROP,
|
40 |
+
TEXT_NAN_CNT, TOKENIZED_FIELD, TOT_OPEN_WORDS,
|
41 |
+
TOT_WORDS, TXT_LEN, VOCAB, WORD, extract_field,
|
42 |
+
load_truncated_dataset)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
43 |
from .embeddings import Embeddings
|
44 |
from .npmi import nPMI
|
45 |
from .zipf import Zipf
|
|
|
138 |
_TOP_N = 100
|
139 |
_CVEC = CountVectorizer(token_pattern="(?u)\\b\\w+\\b", lowercase=True)
|
140 |
|
141 |
+
|
142 |
class DatasetStatisticsCacheClass:
|
143 |
def __init__(
|
144 |
self,
|
|
|
237 |
# path to the directory used for caching
|
238 |
if not isinstance(text_field, str):
|
239 |
text_field = "-".join(text_field)
|
240 |
+
# if isinstance(label_field, str):
|
241 |
# label_field = label_field
|
242 |
+
# else:
|
243 |
# label_field = "-".join(label_field)
|
244 |
self.cache_path = pjoin(
|
245 |
self.cache_dir,
|
246 |
+
f"{dset_name}_{dset_config}_{split_name}_{text_field}", # {label_field},
|
247 |
)
|
248 |
if not isdir(self.cache_path):
|
249 |
logs.warning("Creating cache directory %s." % self.cache_path)
|
|
|
272 |
# Needed for UI
|
273 |
self.dup_counts_df_fid = pjoin(self.cache_path, "dup_counts_df.feather")
|
274 |
# Needed for UI
|
275 |
+
self.fig_tok_length_fid = pjoin(self.cache_path, "fig_tok_length.png")
|
276 |
|
277 |
## General text stats
|
278 |
# Needed for UI
|
279 |
self.general_stats_json_fid = pjoin(self.cache_path, "general_stats_dict.json")
|
280 |
# Needed for UI
|
281 |
+
self.sorted_top_vocab_df_fid = pjoin(
|
282 |
+
self.cache_path, "sorted_top_vocab.feather"
|
283 |
+
)
|
284 |
## Zipf cache files
|
285 |
# Needed for UI
|
286 |
self.zipf_fid = pjoin(self.cache_path, "zipf_basic_stats.json")
|
|
|
292 |
self.node_list_fid = pjoin(self.cache_path, "node_list.th")
|
293 |
# Needed for UI
|
294 |
self.fig_tree_json_fid = pjoin(self.cache_path, "fig_tree.json")
|
|
|
295 |
|
296 |
self.live = False
|
297 |
|
|
|
331 |
and exists(self.dup_counts_df_fid)
|
332 |
and exists(self.sorted_top_vocab_df_fid)
|
333 |
):
|
334 |
+
logs.info("Loading cached general stats")
|
335 |
self.load_general_stats()
|
336 |
else:
|
337 |
if not self.live:
|
338 |
+
logs.info("Preparing general stats")
|
339 |
self.prepare_general_stats()
|
340 |
if save:
|
341 |
write_df(self.sorted_top_vocab_df, self.sorted_top_vocab_df_fid)
|
342 |
write_df(self.dup_counts_df, self.dup_counts_df_fid)
|
343 |
write_json(self.general_stats_dict, self.general_stats_json_fid)
|
344 |
|
|
|
345 |
def load_or_prepare_text_lengths(self, save=True):
|
346 |
"""
|
347 |
The text length widget relies on this function, which provides
|
|
|
353 |
|
354 |
"""
|
355 |
# Text length figure
|
356 |
+
if self.use_cache and exists(self.fig_tok_length_fid):
|
357 |
self.fig_tok_length_png = mpimg.imread(self.fig_tok_length_fid)
|
|
|
358 |
else:
|
359 |
if not self.live:
|
360 |
self.prepare_fig_text_lengths()
|
361 |
if save:
|
362 |
+
self.fig_tok_length.savefig(self.fig_tok_length_fid)
|
|
|
363 |
# Text length dataframe
|
364 |
if self.use_cache and exists(self.length_df_fid):
|
365 |
self.length_df = feather.read_feather(self.length_df_fid)
|
|
|
386 |
if not self.live:
|
387 |
if self.tokenized_df is None:
|
388 |
self.tokenized_df = self.do_tokenization()
|
389 |
+
self.tokenized_df[LENGTH_FIELD] = self.tokenized_df[TOKENIZED_FIELD].apply(
|
390 |
+
len
|
|
|
|
|
|
|
391 |
)
|
392 |
+
self.length_df = self.tokenized_df[
|
393 |
+
[LENGTH_FIELD, OUR_TEXT_FIELD]
|
394 |
+
].sort_values(by=[LENGTH_FIELD], ascending=True)
|
395 |
|
396 |
def prepare_text_length_stats(self):
|
397 |
if not self.live:
|
398 |
+
if (
|
399 |
+
self.tokenized_df is None
|
400 |
+
or LENGTH_FIELD not in self.tokenized_df.columns
|
401 |
+
or self.length_df is None
|
402 |
+
):
|
403 |
self.prepare_length_df()
|
404 |
+
avg_length = sum(self.tokenized_df[LENGTH_FIELD]) / len(
|
405 |
+
self.tokenized_df[LENGTH_FIELD]
|
406 |
+
)
|
407 |
self.avg_length = round(avg_length, 1)
|
408 |
std_length = statistics.stdev(self.tokenized_df[LENGTH_FIELD])
|
409 |
self.std_length = round(std_length, 1)
|
410 |
self.num_uniq_lengths = len(self.length_df["length"].unique())
|
411 |
+
self.length_stats_dict = {
|
412 |
+
"avg length": self.avg_length,
|
413 |
+
"std length": self.std_length,
|
414 |
+
"num lengths": self.num_uniq_lengths,
|
415 |
+
}
|
416 |
|
417 |
def prepare_fig_text_lengths(self):
|
418 |
if not self.live:
|
419 |
+
if (
|
420 |
+
self.tokenized_df is None
|
421 |
+
or LENGTH_FIELD not in self.tokenized_df.columns
|
422 |
+
):
|
423 |
self.prepare_length_df()
|
424 |
self.fig_tok_length = make_fig_lengths(self.tokenized_df, LENGTH_FIELD)
|
425 |
|
426 |
+
def load_or_prepare_embeddings(self):
|
427 |
+
self.embeddings = Embeddings(self, use_cache=self.use_cache)
|
428 |
+
self.embeddings.make_hierarchical_clustering()
|
429 |
+
self.node_list = self.embeddings.node_list
|
430 |
+
self.fig_tree = self.embeddings.fig_tree
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
431 |
|
432 |
# get vocab with word counts
|
433 |
def load_or_prepare_vocab(self, save=True):
|
|
|
437 |
:param
|
438 |
:return:
|
439 |
"""
|
440 |
+
if self.use_cache and exists(self.vocab_counts_df_fid):
|
|
|
|
|
|
|
441 |
logs.info("Reading vocab from cache")
|
442 |
self.load_vocab()
|
443 |
self.vocab_counts_filtered_df = filter_vocab(self.vocab_counts_df)
|
|
|
484 |
write_df(self.dup_counts_df, self.dup_counts_df_fid)
|
485 |
|
486 |
def load_general_stats(self):
|
487 |
+
self.general_stats_dict = json.load(
|
488 |
+
open(self.general_stats_json_fid, encoding="utf-8")
|
489 |
+
)
|
490 |
with open(self.sorted_top_vocab_df_fid, "rb") as f:
|
491 |
self.sorted_top_vocab_df = feather.read_feather(f)
|
492 |
self.text_nan_count = self.general_stats_dict[TEXT_NAN_CNT]
|
|
|
521 |
if not self.live:
|
522 |
if self.tokenized_df is None:
|
523 |
self.load_or_prepare_tokenized_df()
|
524 |
+
dup_df = self.tokenized_df[self.tokenized_df.duplicated([OUR_TEXT_FIELD])]
|
|
|
525 |
self.dup_counts_df = pd.DataFrame(
|
526 |
dup_df.pivot_table(
|
527 |
columns=[OUR_TEXT_FIELD], aggfunc="size"
|
|
|
561 |
write_json({"dset peek": self.dset_peek}, self.dset_peek_json_fid)
|
562 |
|
563 |
def load_or_prepare_tokenized_df(self, save=True):
|
564 |
+
if self.use_cache and exists(self.tokenized_df_fid):
|
565 |
self.tokenized_df = feather.read_feather(self.tokenized_df_fid)
|
566 |
else:
|
567 |
if not self.live:
|
|
|
573 |
write_df(self.tokenized_df, self.tokenized_df_fid)
|
574 |
|
575 |
def load_or_prepare_text_dset(self, save=True):
|
576 |
+
if self.use_cache and exists(self.text_dset_fid):
|
577 |
# load extracted text
|
578 |
self.text_dset = load_from_disk(self.text_dset_fid)
|
579 |
logs.warning("Loaded dataset from disk")
|
|
|
691 |
zipf_dict = json.load(f)
|
692 |
self.z = Zipf()
|
693 |
self.z.load(zipf_dict)
|
|
|
|
|
694 |
self.zipf_fig = read_plotly(self.zipf_fig_fid)
|
695 |
elif self.use_cache and exists(self.zipf_fid):
|
696 |
# TODO: Read zipf data so that the vocab is there.
|
|
|
753 |
and exists(self.npmi_terms_fid)
|
754 |
and json.load(open(self.npmi_terms_fid))["available terms"] != []
|
755 |
):
|
756 |
+
available_terms = json.load(open(self.npmi_terms_fid))["available terms"]
|
757 |
else:
|
758 |
+
true_false = [
|
759 |
+
term in self.dstats.vocab_counts_df.index for term in self.termlist
|
760 |
+
]
|
761 |
+
word_list_tmp = [x for x, y in zip(self.termlist, true_false) if y]
|
762 |
+
true_false_counts = [
|
763 |
+
self.dstats.vocab_counts_df.loc[word, CNT] >= self.min_vocab_count
|
764 |
+
for word in word_list_tmp
|
765 |
+
]
|
766 |
+
available_terms = [
|
767 |
+
word for word, y in zip(word_list_tmp, true_false_counts) if y
|
768 |
+
]
|
769 |
+
logs.info(available_terms)
|
770 |
+
with open(self.npmi_terms_fid, "w+") as f:
|
771 |
+
json.dump({"available terms": available_terms}, f)
|
772 |
+
self.available_terms = available_terms
|
773 |
+
return available_terms
|
774 |
+
|
775 |
+
def load_or_prepare_joint_npmi(self, subgroup_pair):
|
|
|
|
|
|
|
|
|
776 |
"""
|
777 |
Run on-the fly, while the app is already open,
|
778 |
as it depends on the subgroup terms that the user chooses
|
|
|
797 |
# When everything is already computed for the selected subgroups.
|
798 |
logs.info("Loading cached joint npmi")
|
799 |
joint_npmi_df = self.load_joint_npmi_df(joint_npmi_fid)
|
800 |
+
npmi_display_cols = [
|
801 |
+
"npmi-bias",
|
802 |
+
subgroup1 + "-npmi",
|
803 |
+
subgroup2 + "-npmi",
|
804 |
+
subgroup1 + "-count",
|
805 |
+
subgroup2 + "-count",
|
806 |
+
]
|
807 |
joint_npmi_df = joint_npmi_df[npmi_display_cols]
|
808 |
# When maybe some things have been computed for the selected subgroups.
|
809 |
else:
|
|
|
812 |
joint_npmi_df, subgroup_dict = self.prepare_joint_npmi_df(
|
813 |
subgroup_pair, subgroup_files
|
814 |
)
|
815 |
+
# Cache new results
|
816 |
+
logs.info("Writing out.")
|
817 |
+
for subgroup in subgroup_pair:
|
818 |
+
write_subgroup_npmi_data(subgroup, subgroup_dict, subgroup_files)
|
819 |
+
with open(joint_npmi_fid, "w+") as f:
|
820 |
+
joint_npmi_df.to_csv(f)
|
|
|
|
|
821 |
else:
|
822 |
joint_npmi_df = pd.DataFrame()
|
823 |
logs.info("The joint npmi df is")
|
|
|
859 |
subgroup_dict[subgroup] = cached_results
|
860 |
logs.info("Calculating for subgroup list")
|
861 |
joint_npmi_df, subgroup_dict = self.do_npmi(subgroup_pair, subgroup_dict)
|
862 |
+
return joint_npmi_df.dropna(), subgroup_dict
|
863 |
|
864 |
# TODO: Update pairwise assumption
|
865 |
def do_npmi(self, subgroup_pair, subgroup_dict):
|
|
|
870 |
:return: Selected identity term's co-occurrence counts with
|
871 |
other words, pmi per word, and nPMI per word.
|
872 |
"""
|
|
|
873 |
logs.info("Initializing npmi class")
|
874 |
npmi_obj = self.set_npmi_obj()
|
875 |
# Canonical ordering used
|
|
|
877 |
# Calculating nPMI statistics
|
878 |
for subgroup in subgroup_pair:
|
879 |
# If the subgroup data is already computed, grab it.
|
880 |
+
# TODO: Should we set idx and column names similarly to how we set them for cached files?
|
|
|
881 |
if subgroup not in subgroup_dict:
|
882 |
logs.info("Calculating statistics for %s" % subgroup)
|
883 |
vocab_cooc_df, pmi_df, npmi_df = npmi_obj.calc_metrics(subgroup)
|
884 |
+
# Store the nPMI information for the current subgroups
|
885 |
+
subgroup_dict[subgroup] = (vocab_cooc_df, pmi_df, npmi_df)
|
886 |
+
# Pair the subgroups together, indexed by all words that
|
887 |
+
# co-occur between them.
|
888 |
+
logs.info("Computing pairwise npmi bias")
|
889 |
+
paired_results = npmi_obj.calc_paired_metrics(subgroup_pair, subgroup_dict)
|
890 |
+
UI_results = make_npmi_fig(paired_results, subgroup_pair)
|
891 |
+
return UI_results, subgroup_dict
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
892 |
|
893 |
def set_npmi_obj(self):
|
894 |
"""
|
|
|
962 |
def get_available_terms(self):
|
963 |
return self.load_or_prepare_npmi_terms()
|
964 |
|
965 |
+
|
966 |
def dummy(doc):
|
967 |
return doc
|
968 |
|
969 |
+
|
970 |
def count_vocab_frequencies(tokenized_df):
|
971 |
"""
|
972 |
Based on an input pandas DataFrame with a 'text' column,
|
|
|
981 |
)
|
982 |
# We do this to calculate per-word statistics
|
983 |
# Fast calculation of single word counts
|
984 |
+
logs.info(
|
985 |
+
"Fitting dummy tokenization to make matrix using the previous tokenization"
|
986 |
+
)
|
987 |
cvec.fit(tokenized_df[TOKENIZED_FIELD])
|
988 |
document_matrix = cvec.transform(tokenized_df[TOKENIZED_FIELD])
|
989 |
batches = np.linspace(0, tokenized_df.shape[0], _NUM_VOCAB_BATCHES).astype(int)
|
|
|
1004 |
word_count_df.index.name = WORD
|
1005 |
return word_count_df
|
1006 |
|
1007 |
+
|
1008 |
def calc_p_word(word_count_df):
|
1009 |
# p(word)
|
1010 |
word_count_df[PROP] = word_count_df[CNT] / float(sum(word_count_df[CNT]))
|
|
|
1015 |
|
1016 |
def filter_vocab(vocab_counts_df):
|
1017 |
# TODO: Add warnings (which words are missing) to log file?
|
1018 |
+
filtered_vocab_counts_df = vocab_counts_df.drop(_CLOSED_CLASS, errors="ignore")
|
|
|
1019 |
filtered_count = filtered_vocab_counts_df[CNT]
|
1020 |
filtered_count_denom = float(sum(filtered_vocab_counts_df[CNT]))
|
1021 |
filtered_vocab_counts_df[PROP] = filtered_count / filtered_count_denom
|
|
|
1024 |
|
1025 |
## Figures ##
|
1026 |
|
1027 |
+
|
1028 |
def write_plotly(fig, fid):
|
1029 |
write_json(plotly.io.to_json(fig), fid)
|
1030 |
|
1031 |
+
|
1032 |
def read_plotly(fid):
|
1033 |
fig = plotly.io.from_json(json.load(open(fid, encoding="utf-8")))
|
1034 |
return fig
|
1035 |
|
1036 |
+
|
1037 |
def make_fig_lengths(tokenized_df, length_field):
|
1038 |
+
fig_tok_length, axs = plt.subplots(figsize=(15, 6), dpi=150)
|
1039 |
+
sns.histplot(data=tokenized_df[length_field], kde=True, bins=100, ax=axs)
|
1040 |
+
sns.rugplot(data=tokenized_df[length_field], ax=axs)
|
1041 |
return fig_tok_length
|
1042 |
|
1043 |
+
|
1044 |
def make_fig_labels(label_df, label_names, label_field):
|
1045 |
labels = label_df[label_field].unique()
|
1046 |
label_sums = [len(label_df[label_df[label_field] == label]) for label in labels]
|
|
|
1121 |
return fig
|
1122 |
|
1123 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1124 |
## Input/Output ###
|
1125 |
|
1126 |
|
|
|
1174 |
with open(json_fid, "w", encoding="utf-8") as f:
|
1175 |
json.dump(json_dict, f)
|
1176 |
|
1177 |
+
|
1178 |
def write_subgroup_npmi_data(subgroup, subgroup_dict, subgroup_files):
|
1179 |
"""
|
1180 |
Saves the calculated nPMI statistics to their output files.
|
|
|
1194 |
with open(subgroup_cooc_fid, "w+") as f:
|
1195 |
subgroup_cooc_df.to_csv(f)
|
1196 |
|
1197 |
+
|
1198 |
def write_zipf_data(z, zipf_fid):
|
1199 |
zipf_dict = {}
|
1200 |
zipf_dict["xmin"] = int(z.xmin)
|
|
|
1206 |
zipf_dict["uniq_ranks"] = [int(rank) for rank in z.uniq_ranks]
|
1207 |
with open(zipf_fid, "w+", encoding="utf-8") as f:
|
1208 |
json.dump(zipf_dict, f)
|
|