Upload folder using huggingface_hub

medqa_corpus_en.py

@@ -40,19 +40,19 @@ Surgery_Schwartz.txt            Surgery                     Clinical medicine
 
 SUBJECT_SUBSETS = {
     "core_clinical":
-        ["Anatomy_Gray.txt", "First_Aid_Step1.txt", "First_Aid_Step2.txt", "Immunology_Janeway.txt",
-         "InternalMed_Harrison.txt", "Neurology_Adams.txt", "Obstentrics_Williams.txt", "Pathoma_Husain.txt",
-         "Pediatrics_Nelson.txt", "Surgery_Schwartz.txt"],
+        ["Anatomy_Gray", "First_Aid_Step1", "First_Aid_Step2", "Immunology_Janeway",
+         "InternalMed_Harrison", "Neurology_Adams", "Obstentrics_Williams", "Pathoma_Husain", "Pediatrics_Nelson",
+         "Surgery_Schwartz"],
     "basic_biology":
-        ["Biochemistry_Lippincott.txt", "Cell_Biology_Alberts.txt", "Histology_Ross.txt", "Pathology_Robbins.txt",
-         "Physiology_Levy.txt"],
+        ["Biochemistry_Lippincott", "Cell_Biology_Alberts", "Histology_Ross", "Pathology_Robbins", "Physiology_Levy"],
     "pharmacology":
-        ["Pharmacology_Katzung.txt"],
+        ["Pharmacology_Katzung"],
     "psychiatry":
-        ["Psichiatry_DSM-5.txt"]
+        ["Psichiatry_DSM-5"]
 }
 
 from pathlib import Path
+import json
 try:
     from ogbujipt.text_helper import text_splitter
 except ImportError as e:
@@ -97,7 +97,7 @@ OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
 SOFTWARE.
 """
 
-_URLS = ["textbooks_en.zip"]
+_URLS = ["textbooks_en_jsonl.zip"]
 
 _CITATION = """\
 @article{jin2021disease,
@@ -112,53 +112,14 @@ _CITATION = """\
 }
 """
 
-REPL_MAP = {
-    '\n': ' ',
-    '\u2019': "'",
-    '\u2013': '/',
-    '\ufb02': 'fl',
-    '\u2014': ' - ',
-    '\u201c': "'",
-    '\u2003': ' ',
-    '\u201d': "'",
-    '\u2193': 'decreased',
-}
-
-
-def recursive_replace(text, items=None):
-    items = items if items is not None else list(REPL_MAP.items())
-    if items:
-        old, new = items[0]
-        return recursive_replace(text.replace(old, new), items[1:])
-    else:
-        return text
-
-
-def proportion_of_ascii_characters(string):
-    # Count the number of ASCII characters in the string.
-    num_ascii_characters = 0
-    for character in string:
-        if ord(character) < 128:
-            num_ascii_characters += 1
-
-    # Calculate the proportion of ASCII characters in the string.
-    prop = num_ascii_characters / len(string)
-
-    return prop
-
 
 def get_med_qa_textbooks(location, subset_name):
-    for textbook_content in location.glob('*.txt'):
-        if textbook_content.name in SUBJECT_SUBSETS[subset_name]:
-            r_splitter = RecursiveCharacterTextSplitter(
-                chunk_size=MAX_CHUNK_SIZE,
-                chunk_overlap=CHUNK_OVERLAP,
-                separators=["\n", " ", ".  ", ""]
-            )
-            for chunk in r_splitter.split_text(textbook_content.open(encoding='utf-8').read()):
-                chunk = recursive_replace(chunk)
-                if proportion_of_ascii_characters(chunk) >= .9:
-                    yield {"text": chunk, "source": str(textbook_content)}
+    for textbook_content in location.glob('*.jsonl'):
+        textbook_name = textbook_content.name.split('.')[0]
+        if textbook_name in SUBJECT_SUBSETS[subset_name]:
+            with textbook_content.open("r") as fid:
+                for line in fid:
+                    yield json.loads(line)
 
 
 # TODO: Name of the dataset usually matches the script name with CamelCase instead of snake_case

textbooks_en_jsonl.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:88badbb560ee3874b2b750225ba772bebfb091569c4c27ac9e220bd4deb6fedd
+size 29196996