implemented preprocessing for secondary subfields

all2023/val.json

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:728f5ae7ff4f3907c9a5e6f3b35d1a2968a02cd825f50dc575efe94c7d4e4ffa
-size 409838242
+oid sha256:d695efd145ac745cf81aab5073fc6938b7562b4330f1aed2cb029e3aae958627
+size 412108395

major/test.json

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c7849149e5f504fd8f99efa3b5f1c2d65b541b2b76640d563bababa4b4f6d395
-size 406464012
+oid sha256:94c2a729e10b65e46a530ee56fe1f17ecf588ffbf978a9a15166a385134e48a6
+size 406488121

major/train.json

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:37aa1843751273b4e27633b1ec1a0fad6256a9143ed48d81bbf34ffae03116c4
-size 1639035113
+oid sha256:9ceb77a8b2c6fd61decbc05c5c469341637638dfe513e4238c994b79c8edc2fc
+size 1639130854

minor/test.json

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:667308dc5ddfa3cf17869673427e5ce85abf054fee301e8a1cf58ee04e6d9d3a
-size 1453222079
+oid sha256:fab45ae4a61a692399a62538566ae755750592529ad21135a03946192e420fcf
+size 1453299801

minor/train.json

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8834e6cfa89d589cffcc5335a3bc2371ce91f6627c4a653ec689b3e370582b8e
-size 5823161950
+oid sha256:1640a3c3cbf951940ed0edab35eac1c5d7c638c7e97f06d12fdd1902f7cb3e71
+size 5823472832

preprocess.sh

@@ -0,0 +1,13 @@
+# # test
+# python preprocess_major_minor.py -d major -s test
+
+# major/minor cats
+for dataset in major minor
+do
+    for split in train test
+    do
+        python preprocess_major_minor.py -d $dataset -s $split
+    done
+done
+# all 2023 corpus
+python preprocess_all2023.py

preprocess_all2023.py

@@ -34,12 +34,9 @@ output_path = args.output_path
 # Remaining imports
 # 
 import pandas as pd
-import numpy as np
 import os
 from tqdm import tqdm
-from arxiv_classifier.utils import (CATEGORY_ALIASES, 
-                                    GENERAL_CATEGORIES, 
-                                    IGNORED_CATEGOREIES)
+from util_preprocess import preprocess_primary_secondary
 
 # %% [markdown]
 # Rename columns to reflect standarized terminology:
@@ -89,10 +86,8 @@ for i, original_df in tqdm(enumerate(pd.read_csv(file_path, sep='\t', chunksize=
         if feature not in original_df.columns:
             original_df[feature] = ""
 
-    # ignore entries from ignored categories (i.e., primary subfield)
-    df = original_df[~original_df['primary_subfield'].isin(GENERAL_CATEGORIES | IGNORED_CATEGOREIES)]
-    if df['primary_subfield'].isin(CATEGORY_ALIASES.keys()).sum():
-        ValueError("IMPLEMENT CODE TO CHANGE CATEGORY_ALIASES ")
+    # Apply preprocessing rules
+    df = preprocess_primary_secondary(original_df)
 
     # convert all columns to string per huggingface requirements
     if i == 0:

preprocess_major_minor.py

@@ -34,12 +34,9 @@ output_path = args.output_path
 # Remaining imports
 # 
 import pandas as pd
-import numpy as np
 import os
 from tqdm import tqdm
-from arxiv_classifier.utils import (CATEGORY_ALIASES, 
-                                    GENERAL_CATEGORIES, 
-                                    IGNORED_CATEGOREIES)
+from util_preprocess import preprocess_primary_secondary
 
 # %% [markdown]
 # Rename columns to reflect standarized terminology:
@@ -67,14 +64,14 @@ for i, original_df in tqdm(enumerate(pd.read_json(file_path, lines=True, chunksi
         'prime_category':   'primary_subfield',
         'abs_categories':   'secondary_subfield'
     }, inplace=True)
-    # ignore entries from ignored categories (i.e., primary subfield)
-    df = original_df[~original_df['field'].isin(GENERAL_CATEGORIES | IGNORED_CATEGOREIES)]
-    if df['primary_subfield'].isin(CATEGORY_ALIASES.keys()).sum():
-        ValueError("IMPLEMENT CODE TO CHANGE CATEGORY_ALIASES ")
 
+    # Apply preprocessing rules
+    df = preprocess_primary_secondary(original_df)
+
+    # Save to file
     if i == 0:
         df.to_json(save_path, lines=True, orient="records")
     else:
         df.to_json(save_path, lines=True, orient="records", mode='a')
-
+print("Saved to file: ", save_path)
 # %%

util_preprocess.py

@@ -0,0 +1,24 @@
+from arxiv_classifier.utils import (CATEGORY_ALIASES, 
+                                    GENERAL_CATEGORIES, 
+                                    IGNORED_CATEGOREIES)
+
+IGNORE = GENERAL_CATEGORIES | IGNORED_CATEGOREIES
+
+def preprocess_primary_secondary(original_df):
+    df = original_df[~original_df['field'].isin(IGNORE)]
+    if df['primary_subfield'].isin(CATEGORY_ALIASES.keys()).sum():
+        ValueError("IMPLEMENT CODE TO CHANGE CATEGORY_ALIASES ")
+
+    def preprocess_secondary(s):
+        # split string to list by space
+        subfields = s.split()
+        # ignore entries from ignored categories (i.e., primary subfield)
+        subfields = [subfield for subfield in subfields if subfield not in IGNORE]
+        # substitute subfield aliases
+        if set(subfields) & set(CATEGORY_ALIASES.keys()):
+            ValueError("IMPLEMENT CODE TO CHANGE CATEGORY_ALIASES ")
+        return subfields
+    
+    # Preprocess secondary subfields
+    df.loc[:, 'secondary_subfield'] = df['secondary_subfield'].apply(preprocess_secondary)
+    return df