Ezi's picture
Upload 312 files
46df0b6
raw
history blame contribute delete
No virus
2.3 kB
import gradio as gr
from widgets.widget_base import Widget
from data_measurements.dataset_statistics import DatasetStatisticsCacheClass as dmt_cls
import utils
import utils.dataset_utils as ds_utils
logs = utils.prepare_logging(__file__)
class Duplicates(Widget):
def __init__(self):
duplicates_text = f"""
Use this widget to identify text strings that appear more than once.
A model's training and testing may be negatively affected by unwarranted duplicates ([Lee et al., 2021](https://arxiv.org/abs/2107.06499))
------
### Here is the list of all the duplicated items and their counts in the dataset.
"""
self.duplicates_intro = gr.Markdown(render=False, value=duplicates_text)
self.duplicates_df = gr.DataFrame(render=False)
self.duplicates_text = gr.Markdown(render=False)
def render(self):
with gr.TabItem(f"Duplicates"):
self.duplicates_intro.render()
self.duplicates_text.render()
self.duplicates_df.render()
def update(self, dstats: dmt_cls):
output = {}
if not dstats.duplicates_results:
output[self.duplicates_df] = gr.DataFrame.update(visible=False)
output[self.duplicates_text] = gr.Markdown.update(visible=True,
value="There are no duplicates in this dataset! 🥳")
else:
dupes_df_tmp = ds_utils.counter_dict_to_df(dstats.dups_dict, key_as_column=True)
dupes_df_tmp.columns = ["instance", "count"]
# Nice to have the counts show up first, because the instances
# can be quite long (and run off the page)
dupes_df = dupes_df_tmp[["count", "instance"]]
output[self.duplicates_df] = gr.DataFrame.update(visible=True, value=dupes_df)
duplicates_text = f"The fraction of data that is duplicate is {str(round(dstats.dups_frac, 4))}"
output[self.duplicates_text] = gr.Markdown.update(value=duplicates_text, visible=True)
return output
@property
def output_components(self):
return [
self.duplicates_text,
self.duplicates_df,
]
def add_events(self, state: gr.State):
pass