raw paper scrapes

1 files changed, 49 insertions, 0 deletions

diff --git a/s2-fetch-pdf.py b/s2-fetch-pdf.py
new file mode 100644
index 00000000..7d834ada
--- /dev/null
+++ b/s2-fetch-pdf.py
@@ -0,0 +1,49 @@
+import os
+import sys
+import csv
+import subprocess
+import time
+import random
+import re
+import json
+import click
+from s2 import SemanticScholarAPI
+from util import *
+
+s2 = SemanticScholarAPI()
+
+@click.command()
+@click.option('--fn', '-i', default='db_paper_pdf.csv', help='Filename of CSV (id, url,)')
+def fetch_pdfs(fn):
+  lines = read_csv(fn, keys=False)
+  for line in lines:
+    paper_id, url = line
+    fetch_pdf(paper_id, url)
+  print("{} papers processed".format(len(lines)))
+
+def fetch_pdf(paper_id, url):
+  os.makedirs(make_pdf_path(paper_id), exist_ok=True)
+  pdf_fn = make_pdf_fn(paper_id)
+  txt_fn = make_txt_fn(paper_id)
+  if os.path.exists(pdf_fn) or os.path.exists(txt_fn):
+    # return read_json(pdf_fn)
+    return
+  size = s2.fetch_file(url, pdf_fn)
+  if size is None:
+    print("{} empty?".format(paper_id))
+    time.sleep(random.randint(5, 10))
+    return None  
+  print("{} {} kb".format(paper_id, int(size / 1024)))
+  time.sleep(random.randint(5, 10))
+  return
+  # return paper
+
+def make_pdf_path(paper_id):
+  return './datasets/s2/pdf/{}/{}'.format(paper_id[0:2], paper_id)
+def make_pdf_fn(paper_id):
+  return './datasets/s2/pdf/{}/{}/paper.pdf'.format(paper_id[0:2], paper_id)
+def make_txt_fn(paper_id):
+  return './datasets/s2/pdf/{}/{}/paper.txt'.format(paper_id[0:2], paper_id)
+
+if __name__ == '__main__':
+  fetch_pdfs()