ag2435 commited on
Commit
2cf1cf5
·
1 Parent(s): 131c65a

added all2023 train-test split with fulltext

Browse files
README.md CHANGED
@@ -16,10 +16,30 @@ configs:
16
  data_files:
17
  - split: val
18
  path: "all2023/val.json"
 
 
 
 
 
 
19
  ---
20
 
21
  # arXiv Classifier Data
22
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
23
  Standardized terminology:
24
  - Field: Bio/cs/physics
25
  - Subfield: Subcategories within each
@@ -38,11 +58,18 @@ Original data: https://www.dropbox.com/scl/fo/wwu0ifghw4sco09g67frb/h?rlkey=6ddg
38
  **Major**: Dataset of papers between 2010 and 2020 (with some pre-2010 papers) with unbalanced primary subfields to better represent the true distribution of primary categories, which is dominated by a few subfields. Note that the distribution of major subfields is still truncated.
39
 
40
  **All 2023**: All papers published in 2023.
 
 
41
 
42
  ![Primary subfield distribution](figures/Primary_subfield_distribution.png)
43
 
44
  ![Secondary subfield distribution](figures/Secondary_subfield_distribution.png)
45
 
 
 
 
 
 
46
  ## Set up dependencies
47
 
48
  Get subfields to ignore and subfield aliases:
 
16
  data_files:
17
  - split: val
18
  path: "all2023/val.json"
19
+ - config_name: all2023_v2
20
+ data_files:
21
+ - split: train
22
+ path: "all2023_v2/train.json"
23
+ - split: test
24
+ path: "all2023_v2/test.json"
25
  ---
26
 
27
  # arXiv Classifier Data
28
 
29
+ Usage:
30
+ ```
31
+ from datasets import load_dataset, DownloadMode
32
+ # download from HuggingFace
33
+ dataset = load_dataset('mlcore/arxiv-classifier', name=<CONFIG NAME>)
34
+ # load from G2
35
+ dataset = load_dataset('/share/nikola/arxiv_classifier/data/arxiv-classifier', name=<CONFIG NAME>)
36
+ ```
37
+ To force the dataset to be re-generated:
38
+ ```
39
+ dataset = load_dataset('/share/nikola/arxiv_classifier/data/arxiv-classifier', name=<CONFIG NAME>, download_mode=DownloadMode.FORCE_REDOWNLOAD)
40
+ ```
41
+ See: https://huggingface.co/docs/datasets/v2.20.0/en/package_reference/builder_classes#datasets.DownloadMode
42
+
43
  Standardized terminology:
44
  - Field: Bio/cs/physics
45
  - Subfield: Subcategories within each
 
58
  **Major**: Dataset of papers between 2010 and 2020 (with some pre-2010 papers) with unbalanced primary subfields to better represent the true distribution of primary categories, which is dominated by a few subfields. Note that the distribution of major subfields is still truncated.
59
 
60
  **All 2023**: All papers published in 2023.
61
+ - Train: papers with date between January and June (inclusive)
62
+ - Test: papers with date between July and December (inclusive)
63
 
64
  ![Primary subfield distribution](figures/Primary_subfield_distribution.png)
65
 
66
  ![Secondary subfield distribution](figures/Secondary_subfield_distribution.png)
67
 
68
+ To generate subfield distribution plots:
69
+ ```
70
+ python plot_subfield_distributions.py --output_path <path_to_save_plots>
71
+ ```
72
+
73
  ## Set up dependencies
74
 
75
  Get subfields to ignore and subfield aliases:
all2023/val.json CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:4186a5287defbe55cb14b48b96a9d42630b623e645f65a65e18bfa12d4f9bfcb
3
- size 13917099843
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7323614e2e540385689dba0e6252fe16c51e47ad6de946d7c2545f563d4290c3
3
+ size 13921074283
all2023_v2/test.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9bf8222a0cbdbc4889512009069292e108477b9ecf5075522a770dcf5137348e
3
+ size 7220291019
all2023_v2/train.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e06a48f7e9bb778b8333b42481989e2b88a09a9affdbef695bede25c76020c3b
3
+ size 6700804113
figures/Primary_subfield_distribution.png CHANGED

Git LFS Details

  • SHA256: 75029a73138f3eeebaccc3635c6e5cdeb928c3c29188c7302d549c3ce62084f2
  • Pointer size: 130 Bytes
  • Size of remote file: 85.1 kB

Git LFS Details

  • SHA256: b127270384980a78f705842875a09db8d773e9aa67cc2893f3f59b325a34e334
  • Pointer size: 130 Bytes
  • Size of remote file: 85.1 kB
figures/Secondary_subfield_distribution.png CHANGED

Git LFS Details

  • SHA256: 588ea037ccabdc38fe67e898dde2b25e231c7dd11b5c6dffc597bb2325f82a09
  • Pointer size: 130 Bytes
  • Size of remote file: 89.4 kB

Git LFS Details

  • SHA256: 07e3c8981528fef0c5fe28a25067dc3cbab59678b2e97ba4269998104858e935
  • Pointer size: 130 Bytes
  • Size of remote file: 89.5 kB
figures/primary_subfield_distribution.csv ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:820f2f0256f5ea60b703d2bface4a6933a28366207fbc62b32c878004261fd79
3
+ size 10666
figures/secondary_subfield_distribution.csv ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0da253fe17779bbff8a61789c9a809aa285569754be282c79729585fccd0ed51
3
+ size 10640
plot_subfield_distributions.py CHANGED
@@ -1,23 +1,45 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  # Script to plot the distribution of primary and secondary subfields in the minor, major, and all2023 datasets.
2
  # Usage:
3
  # python plot_subfield_distributions.py --output_path <path_to_save_plots>
4
 
 
5
  from argparse import ArgumentParser
6
- parser = ArgumentParser()
7
  # parser.add_argument("--dataset", "-d", type=str, default="minor")
8
  # parser.add_argument("--split", "-s", type=str, default="val")
9
- parser.add_argument("--output_path", "-op", type=str, default="./")
10
- args = parser.parse_args()
11
- output_path = args.output_path
 
12
 
 
13
  #
14
  # Remaining imports
15
  #
16
  import os
 
17
  import plotly.graph_objects as go
18
  from datasets import load_dataset
19
  import itertools
20
-
 
21
  # Get save path
22
  save_path = os.path.join(
23
  output_path,
@@ -26,11 +48,13 @@ save_path = os.path.join(
26
  os.makedirs(save_path, exist_ok=True)
27
  print(f"Saving figures to {save_path}")
28
 
 
29
  # Loading dataset and tokenizer as well
30
  minor_dataset = load_dataset("mlcore/arxiv-classifier")
31
  major_dataset = load_dataset("mlcore/arxiv-classifier", "major")
32
  all2023_dataset = load_dataset("mlcore/arxiv-classifier", "all2023")
33
 
 
34
  # Extracting the primary and secondary subfield distributions
35
  datasets = {
36
  "minor dataset" : minor_dataset['train'],
@@ -43,6 +67,7 @@ secondary_subfield_distributions = {
43
  for name, dataset in datasets.items()
44
  }
45
 
 
46
  # Plotting the two distributions
47
  hists = []
48
  n = min([len(i) for i in primary_subfield_distributions.values()])
@@ -50,8 +75,10 @@ for name, primary_subfield_distribution in primary_subfield_distributions.items(
50
  # primary_subfield_distribution = random.sample(primary_subfield_distribution, n)
51
  hists.append(go.Histogram(x=primary_subfield_distribution, name=name, opacity=0.75, histnorm='probability'))
52
 
 
53
  fig = go.Figure(data=hists)
54
 
 
55
  # Update layout for better readability
56
  fig.update_layout(
57
  height=700,
@@ -71,11 +98,29 @@ fig.update_layout(
71
  borderwidth=1
72
  )
73
  )
74
-
 
75
  # Show the figure
76
  # fig.show()
77
  fig.write_image(os.path.join(save_path,"Primary_subfield_distribution.png"))
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
78
 
 
79
  # Plotting the two distributions
80
  hists = []
81
  n = min([len(i) for i in secondary_subfield_distributions.values() if len(i) > 0])
@@ -83,8 +128,10 @@ for name, secondary_subfield_distribution in secondary_subfield_distributions.it
83
  # secondary_subfield_distribution = random.sample(secondary_subfield_distribution, min(n, len(secondary_subfield_distribution)))
84
  hists.append(go.Histogram(x=secondary_subfield_distribution, name=name, opacity=0.75, histnorm='probability'))
85
 
 
86
  fig = go.Figure(data=hists)
87
 
 
88
  # Update layout for better readability
89
  fig.update_layout(
90
  height=700,
@@ -104,7 +151,24 @@ fig.update_layout(
104
  borderwidth=1
105
  )
106
  )
107
-
 
108
  # Show the figure
109
  # fig.show()
110
- fig.write_image(os.path.join(save_path,"Secondary_subfield_distribution.png"))
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # ---
2
+ # jupyter:
3
+ # jupytext:
4
+ # cell_metadata_filter: -all
5
+ # custom_cell_magics: kql
6
+ # text_representation:
7
+ # extension: .py
8
+ # format_name: percent
9
+ # format_version: '1.3'
10
+ # jupytext_version: 1.11.2
11
+ # kernelspec:
12
+ # display_name: Python 3
13
+ # language: python
14
+ # name: python3
15
+ # ---
16
+
17
+ # %% [markdown]
18
  # Script to plot the distribution of primary and secondary subfields in the minor, major, and all2023 datasets.
19
  # Usage:
20
  # python plot_subfield_distributions.py --output_path <path_to_save_plots>
21
 
22
+ # %%
23
  from argparse import ArgumentParser
24
+ # parser = ArgumentParser()
25
  # parser.add_argument("--dataset", "-d", type=str, default="minor")
26
  # parser.add_argument("--split", "-s", type=str, default="val")
27
+ # parser.add_argument("--output_path", "-op", type=str, default="./")
28
+ # args = parser.parse_args()
29
+ # output_path = args.output_path
30
+ output_path = './'
31
 
32
+ # %%
33
  #
34
  # Remaining imports
35
  #
36
  import os
37
+ import pandas as pd
38
  import plotly.graph_objects as go
39
  from datasets import load_dataset
40
  import itertools
41
+ from datetime import datetime
42
+ # %%
43
  # Get save path
44
  save_path = os.path.join(
45
  output_path,
 
48
  os.makedirs(save_path, exist_ok=True)
49
  print(f"Saving figures to {save_path}")
50
 
51
+ # %%
52
  # Loading dataset and tokenizer as well
53
  minor_dataset = load_dataset("mlcore/arxiv-classifier")
54
  major_dataset = load_dataset("mlcore/arxiv-classifier", "major")
55
  all2023_dataset = load_dataset("mlcore/arxiv-classifier", "all2023")
56
 
57
+ # %%
58
  # Extracting the primary and secondary subfield distributions
59
  datasets = {
60
  "minor dataset" : minor_dataset['train'],
 
67
  for name, dataset in datasets.items()
68
  }
69
 
70
+ # %%
71
  # Plotting the two distributions
72
  hists = []
73
  n = min([len(i) for i in primary_subfield_distributions.values()])
 
75
  # primary_subfield_distribution = random.sample(primary_subfield_distribution, n)
76
  hists.append(go.Histogram(x=primary_subfield_distribution, name=name, opacity=0.75, histnorm='probability'))
77
 
78
+ # %%
79
  fig = go.Figure(data=hists)
80
 
81
+ # %%
82
  # Update layout for better readability
83
  fig.update_layout(
84
  height=700,
 
98
  borderwidth=1
99
  )
100
  )
101
+
102
+ # %%
103
  # Show the figure
104
  # fig.show()
105
  fig.write_image(os.path.join(save_path,"Primary_subfield_distribution.png"))
106
+ # also save distributions as csv
107
+ df_list = []
108
+ for name, primary_subfield_distribution in primary_subfield_distributions.items():
109
+ df = pd.DataFrame(data={
110
+ "primary_subfield": primary_subfield_distribution,
111
+ "dataset": name,
112
+ "count": 1
113
+ })
114
+ df_list.append(df)
115
+ df = pd.concat(df_list)
116
+ counts_by_subfield = df.groupby(['dataset', 'primary_subfield']).agg('count').unstack(level=0)
117
+ counts_by_subfield.columns = counts_by_subfield.columns.droplevel()
118
+ # normalize counts
119
+ counts_by_subfield = counts_by_subfield.div(counts_by_subfield.sum(axis=0), axis=1)
120
+ # print(counts_by_subfield)
121
+ counts_by_subfield.to_csv(os.path.join(save_path, "primary_subfield_distribution.csv"))
122
 
123
+ # %%
124
  # Plotting the two distributions
125
  hists = []
126
  n = min([len(i) for i in secondary_subfield_distributions.values() if len(i) > 0])
 
128
  # secondary_subfield_distribution = random.sample(secondary_subfield_distribution, min(n, len(secondary_subfield_distribution)))
129
  hists.append(go.Histogram(x=secondary_subfield_distribution, name=name, opacity=0.75, histnorm='probability'))
130
 
131
+ # %%
132
  fig = go.Figure(data=hists)
133
 
134
+ # %%
135
  # Update layout for better readability
136
  fig.update_layout(
137
  height=700,
 
151
  borderwidth=1
152
  )
153
  )
154
+
155
+ # %%
156
  # Show the figure
157
  # fig.show()
158
+ fig.write_image(os.path.join(save_path,"Secondary_subfield_distribution.png"))
159
+ # also save distributions as csv
160
+ df_list = []
161
+ for name, secondary_subfield_distribution in secondary_subfield_distributions.items():
162
+ df = pd.DataFrame(data={
163
+ "secondary_subfield": secondary_subfield_distribution,
164
+ "dataset": name,
165
+ "count": 1
166
+ })
167
+ df_list.append(df)
168
+ df = pd.concat(df_list)
169
+ counts_by_subfield = df.groupby(['dataset', 'secondary_subfield']).agg('count').unstack(level=0)
170
+ counts_by_subfield.columns = counts_by_subfield.columns.droplevel()
171
+ # normalize counts
172
+ counts_by_subfield = counts_by_subfield.div(counts_by_subfield.sum(axis=0), axis=1)
173
+ # print(counts_by_subfield)
174
+ counts_by_subfield.to_csv(os.path.join(save_path, "secondary_subfield_distribution.csv"))
preprocess_all2023_v2.py CHANGED
@@ -30,7 +30,7 @@ parser.add_argument("--input_path", "-ip", type=str, default="./data",
30
  args, opt = parser.parse_known_args()
31
 
32
  dataset = 'all2023' # args.dataset
33
- split = 'val' # args.split
34
  output_path = args.output_path
35
  input_path = args.input_path
36
 
@@ -42,6 +42,7 @@ import pandas as pd
42
  import os
43
  import sys
44
  from tqdm import tqdm
 
45
  from util_preprocess import preprocess_primary_secondary
46
 
47
  # %% [markdown]
@@ -57,7 +58,6 @@ from util_preprocess import preprocess_primary_secondary
57
  # - Major category = field
58
 
59
  # %%
60
- # file_path = os.path.join('data', 'raw', dataset, f"{split}_{dataset}_cats_full.json")
61
  file_path = os.path.join(
62
  input_path,
63
  'raw',
@@ -66,7 +66,8 @@ file_path = os.path.join(
66
  )
67
  save_dir = os.path.join(output_path, dataset)
68
  os.makedirs(save_dir, exist_ok=True)
69
- save_path = os.path.join(save_dir, f"{split}.json")
 
70
  print("Reading from file: ", file_path)
71
 
72
  FEATURES = [
@@ -81,6 +82,17 @@ FEATURES = [
81
  'field',
82
  'fulltext'
83
  ]
 
 
 
 
 
 
 
 
 
 
 
84
 
85
  def get_fulltext(row):
86
  filename = f"{row['paper_id']}v{row['version']}.txt"
@@ -105,6 +117,10 @@ count_hasfulltext = []
105
  # Ignore all quotes by setting quoting=3 (corresponds to csv.QUOTE_NONE)
106
  for i, original_df in tqdm(enumerate(pd.read_csv(file_path, sep='\t', chunksize=10, dtype=str, quoting=3))):
107
  count_orig.append( len(original_df) )
 
 
 
 
108
  # set non-existent features to empty string
109
  for feature in FEATURES:
110
  if feature not in original_df.columns:
@@ -113,7 +129,7 @@ for i, original_df in tqdm(enumerate(pd.read_csv(file_path, sep='\t', chunksize=
113
  original_df.loc[:, 'primary_subfield'] = original_df['categories'].apply(lambda x : x.split()[0])
114
  # keep the secondary_subfield a string, will be converted to list in preprocess_primary_secondary()
115
  original_df.loc[:, 'secondary_subfield'] = original_df['categories']
116
- # Apply preprocessing rules
117
  df = preprocess_primary_secondary(original_df)
118
 
119
  count_isvalid.append( len(df) )
@@ -125,13 +141,23 @@ for i, original_df in tqdm(enumerate(pd.read_csv(file_path, sep='\t', chunksize=
125
  df = df.dropna(subset=['fulltext'])
126
  count_hasfulltext.append( len(df) )
127
 
 
 
 
128
  # convert all columns to string per huggingface requirements
 
129
  if i == 0:
130
- df.to_json(save_path, lines=True, orient="records")
 
 
 
131
  else:
132
- df.to_json(save_path, lines=True, orient="records", mode='a')
 
 
 
133
 
134
- print("Saved to: ", save_path)
135
  df_err = pd.DataFrame({
136
  'orig': count_orig,
137
  'isvalid': count_isvalid,
 
30
  args, opt = parser.parse_known_args()
31
 
32
  dataset = 'all2023' # args.dataset
33
+ # split = 'val' # args.split
34
  output_path = args.output_path
35
  input_path = args.input_path
36
 
 
42
  import os
43
  import sys
44
  from tqdm import tqdm
45
+ from datetime import datetime
46
  from util_preprocess import preprocess_primary_secondary
47
 
48
  # %% [markdown]
 
58
  # - Major category = field
59
 
60
  # %%
 
61
  file_path = os.path.join(
62
  input_path,
63
  'raw',
 
66
  )
67
  save_dir = os.path.join(output_path, dataset)
68
  os.makedirs(save_dir, exist_ok=True)
69
+ train_save_path = os.path.join(save_dir, "train.json")
70
+ test_save_path = os.path.join(save_dir, "test.json")
71
  print("Reading from file: ", file_path)
72
 
73
  FEATURES = [
 
82
  'field',
83
  'fulltext'
84
  ]
85
+ def get_date(paper):
86
+ # first get the arxiv id
87
+ paper_id = paper['paper_id']
88
+ # yy is the first two characters of the arxiv id
89
+ yy = paper_id[:2]
90
+ # mm is the next two characters of the arxiv id
91
+ mm = paper_id[2:4]
92
+ # return datatime object
93
+ date_obj = datetime(year=int(f"20{yy}"), month=int(mm), day=1)
94
+ # return date as string in the format YYYY-MM-DDTHH:MM:SS
95
+ return date_obj #.strftime('%Y-%m-%dT%H:%M:%S')
96
 
97
  def get_fulltext(row):
98
  filename = f"{row['paper_id']}v{row['version']}.txt"
 
117
  # Ignore all quotes by setting quoting=3 (corresponds to csv.QUOTE_NONE)
118
  for i, original_df in tqdm(enumerate(pd.read_csv(file_path, sep='\t', chunksize=10, dtype=str, quoting=3))):
119
  count_orig.append( len(original_df) )
120
+ # set `yymm` column
121
+ original_df.loc[:, 'yymm'] = original_df['paper_id'].apply(lambda x: x[:4])
122
+ # set `created` column using yymm from the paper_id as a proxy
123
+ original_df.loc[:, 'created'] = original_df.apply(lambda x: get_date(x), axis=1)
124
  # set non-existent features to empty string
125
  for feature in FEATURES:
126
  if feature not in original_df.columns:
 
129
  original_df.loc[:, 'primary_subfield'] = original_df['categories'].apply(lambda x : x.split()[0])
130
  # keep the secondary_subfield a string, will be converted to list in preprocess_primary_secondary()
131
  original_df.loc[:, 'secondary_subfield'] = original_df['categories']
132
+ # Apply preprocessing rules and add `field` column
133
  df = preprocess_primary_secondary(original_df)
134
 
135
  count_isvalid.append( len(df) )
 
141
  df = df.dropna(subset=['fulltext'])
142
  count_hasfulltext.append( len(df) )
143
 
144
+ # assert that df has all the required columns
145
+ assert all([feature in df.columns for feature in FEATURES]), f"Missing features in df: {FEATURES}"
146
+
147
  # convert all columns to string per huggingface requirements
148
+ DATE_CUTOFF = datetime(2023, 7, 1)
149
  if i == 0:
150
+ # save the rows with created date before 2023-06-01 to train.json
151
+ df[df['created'] < DATE_CUTOFF].to_json(train_save_path, lines=True, orient="records") #, date_format='iso')
152
+ # save the rows with created date after 2023-06-01 to test.json
153
+ df[df['created'] >= DATE_CUTOFF].to_json(test_save_path, lines=True, orient="records") #, date_format='iso')
154
  else:
155
+ # save the rows with created date before 2023-06-01 to train.json
156
+ df[df['created'] < DATE_CUTOFF].to_json(train_save_path, lines=True, orient="records", mode='a')#, date_format='iso')
157
+ # save the rows with created date after 2023-06-01 to test.json
158
+ df[df['created'] >= DATE_CUTOFF].to_json(test_save_path, lines=True, orient="records", mode='a')#, date_format='iso')
159
 
160
+ # print("Saved to: ", save_path)
161
  df_err = pd.DataFrame({
162
  'orig': count_orig,
163
  'isvalid': count_isvalid,
util_preprocess.py CHANGED
@@ -6,6 +6,17 @@ from arxiv_classifier.utils.maps import (CATEGORY_ALIASES,
6
  IGNORE = GENERAL_CATEGORIES | IGNORED_CATEGOREIES
7
 
8
  def preprocess_primary_secondary(df):
 
 
 
 
 
 
 
 
 
 
 
9
  def substitute_aliases(s):
10
  if s in CATEGORY_ALIASES:
11
  return CATEGORY_ALIASES[s]
@@ -38,6 +49,8 @@ def preprocess_primary_secondary(df):
38
  # Preprocess secondary subfields
39
  df.loc[:, 'secondary_subfield'] = df['secondary_subfield'].apply(preprocess_secondary)
40
 
 
 
41
  # tests
42
  # assert that each primary subfield is in SUBFIELD_MAP
43
  assert all(df['primary_subfield'].isin(SUBFIELD_MAP.keys())), df[~df['primary_subfield'].isin(SUBFIELD_MAP.keys())]
 
6
  IGNORE = GENERAL_CATEGORIES | IGNORED_CATEGOREIES
7
 
8
  def preprocess_primary_secondary(df):
9
+ """
10
+ Applies substitutions and filtering to primary and secondary subfields in the DataFrame.
11
+ Adds a new column `field` to the DataFrame based on the primary subfield.
12
+ (If there exists a `field` column in the original DataFrame, it will be overwritten.)
13
+
14
+ Args:
15
+ df (pd.DataFrame): DataFrame with columns `primary_subfield` and `secondary_subfield`
16
+
17
+ Returns:
18
+ pd.DataFrame: DataFrame with columns `primary_subfield`, `secondary_subfield`, and `field`
19
+ """
20
  def substitute_aliases(s):
21
  if s in CATEGORY_ALIASES:
22
  return CATEGORY_ALIASES[s]
 
49
  # Preprocess secondary subfields
50
  df.loc[:, 'secondary_subfield'] = df['secondary_subfield'].apply(preprocess_secondary)
51
 
52
+ # Generate new column `field` from primary subfield
53
+ df.loc[:, 'field'] = df['primary_subfield'].apply(lambda x: x.split('.')[0])
54
  # tests
55
  # assert that each primary subfield is in SUBFIELD_MAP
56
  assert all(df['primary_subfield'].isin(SUBFIELD_MAP.keys())), df[~df['primary_subfield'].isin(SUBFIELD_MAP.keys())]