lvwerra HF Staff commited on
Commit
9223ca1
·
1 Parent(s): a863009

add filters

Browse files
Files changed (1) hide show
  1. app.py +12 -5
app.py CHANGED
@@ -34,23 +34,30 @@ def parse_notebook(filter_options):
34
  reset_tmp_folder()
35
 
36
  found_notebook = False
37
- counter = 0
38
  while not found_notebook:
39
- counter+=1
40
- print(counter, filter_options)
41
  notebook_data = next(ds_iter)
42
  notebook_string = notebook_data["text"]
43
  notebook_id = notebook_data["id"].split("/")[-1]
44
 
45
  if filter_options == "none":
46
  found_notebook = True
 
 
 
47
  elif filter_options == ">1MB":
48
  if len(notebook_string)>1 * 1024 * 1024:
49
  found_notebook = True
 
 
 
 
 
 
 
50
  elif filter_options == "has outputs":
51
  notebook_parsed = nbformat.reads(notebook_string, as_version=4)
52
  (notebook_body, resources) = html_exporter.from_notebook_node(notebook_parsed)
53
- if len(resources)>0:
54
  found_notebook = True
55
 
56
  out_path = os.path.join(TMP_DIR, notebook_id)
@@ -68,7 +75,7 @@ def parse_notebook(filter_options):
68
 
69
  with gr.Blocks() as demo:
70
  gr.Markdown("# Kaggle Notebooks")
71
- filter_options = gr.Radio(["none", "has outputs", ">1MB"], value="none", label="Notebook filters", info="A lot of notebooks are short or have the outputs stripped - filters help finding interesting ones."),
72
 
73
  button = gr.Button("Show next!")
74
  file = gr.File()
 
34
  reset_tmp_folder()
35
 
36
  found_notebook = False
 
37
  while not found_notebook:
 
 
38
  notebook_data = next(ds_iter)
39
  notebook_string = notebook_data["text"]
40
  notebook_id = notebook_data["id"].split("/")[-1]
41
 
42
  if filter_options == "none":
43
  found_notebook = True
44
+ elif filter_options == ">10MB":
45
+ if len(notebook_string)>10 * 1024 * 1024:
46
+ found_notebook = True
47
  elif filter_options == ">1MB":
48
  if len(notebook_string)>1 * 1024 * 1024:
49
  found_notebook = True
50
+ elif filter_options == ">100KB":
51
+ if len(notebook_string)>100 * 1024:
52
+ found_notebook = True
53
+ elif filter_options == ">10KB":
54
+ if len(notebook_string)>10 * 1024:
55
+ found_notebook = True
56
+ # too slow
57
  elif filter_options == "has outputs":
58
  notebook_parsed = nbformat.reads(notebook_string, as_version=4)
59
  (notebook_body, resources) = html_exporter.from_notebook_node(notebook_parsed)
60
+ if len(resources["outputs"])>0:
61
  found_notebook = True
62
 
63
  out_path = os.path.join(TMP_DIR, notebook_id)
 
75
 
76
  with gr.Blocks() as demo:
77
  gr.Markdown("# Kaggle Notebooks")
78
+ filter_options = gr.Radio(["none",">10KB", ">100KB", ">1MB", ">10MB"], value="none", label="Notebook filters", info="A lot of notebooks are short or have the outputs stripped - filters help finding interesting ones."),
79
 
80
  button = gr.Button("Show next!")
81
  file = gr.File()