s2 scripts

1 files changed, 39 insertions, 0 deletions

diff --git a/check-counts.py b/check-counts.py
new file mode 100644
index 00000000..4fed4494
--- /dev/null
+++ b/check-counts.py
@@ -0,0 +1,39 @@
+import os
+import sys
+import csv
+from math import ceil
+import subprocess
+import random
+
+import click
+
+@click.command()
+def check_counts():
+  """Split a CSV into groups."""
+  mypath = './datasets/scholar/entries/'
+  onlyfiles = [f for f in os.listdir(mypath) if os.path.isfile(os.path.join(mypath, f))]
+  recs = []
+  for f in onlyfiles:
+    with open(os.path.join(mypath, f), 'rb') as f:
+      reader = csv.reader(f, delimiter='|')
+      print f
+      print repr(reader)
+      lines = list(reader)
+      rec = lines[0]
+      recs.append(rec)
+
+  out_fn = './datasets/scholar_entries.csv'
+  write_csv(out_fn, keys=None, chunk=recs)
+
+# Write a CSV
+def write_csv(fn, keys, chunk):
+  print(fn)
+  with open(fn, 'w') as f:
+    writer = csv.writer(f)
+    if keys is not None:
+      writer.writerow(keys)
+    for row in chunk:
+      writer.writerow(row)
+
+if __name__ == '__main__':
+  check_counts()