Pandas operations Demo using PyScript

Demo developed by Dr. Tirthajyoti sarkar

Check out the announcement from Anaconda about PyScript framework

PyScript is awesome, it just takes a few seconds to load...

To provide the power of Python on your browser, this technology needs to load a special virtual machine and all necessary Python modules

So, sit back and relax. We are loading the modules and generating some random variables as a test. You can start interacting with the page when you see a histogram of those random variates displayed below.

We promise that, once loaded, your interactions will be Python-powered and fast! You can also read more about this awesome framework here at Anaconda's engineering blog.

Anaconda Engineering Blog about PyScript

Test: Random integers

PyScript loading and initializing...

import matplotlib.pyplot as plt import numpy as np integers = np.random.randint(0,100,100) fig1,ax1 = plt.subplots(1,1,dpi=150,figsize=(5,2)) ax1.hist(integers,bins=20,edgecolor='k',color='blue') ax1.set_title("Histogram of random integers",fontsize=11) pyscript.write('test-plot',fig1)

Pandas - random dataframe

Let's generate a Pandas DataFrame with some random nunbers (drawn from the ubiquitous Gaussian Normal distribution).

The pretty formatting of the table is done, of course, using CSS. Check the style section. The Pandas pd.to_html() function has a classes argument that accepts a CSS class or a list of classes. We utilize that formatting this table.

Interested in the statistics of this dataset?

This code is non-trivial as the randomized generation of the DataFrame with the button above makes it imperative that we read the exact content shown above. So, we use the document.getElementbyId() function to accomplish this and read back that HTML into Pandas using the pd.read_html() function.

Download file and analyze

Now, we will download a CSV file from an externally hosted URL (a GitHub repo) and show its contents and statistics.

First 5 records of the downloaded file

Statistics of the downloaded file

import pandas as pd, numpy as np from js import document from pyodide.http import open_url def random_table(n_cols=5,n_rows=20,*ags, **kws): cols = ['Col-'+str(i) for i in range(1,n_cols+1)] idx = ['Row-'+str(i) for i in range(1,n_rows+1)] data = np.round(np.random.normal(size=(n_rows,n_cols)),2) df = pd.DataFrame(data,columns=cols,index=idx) return df def show_table(*ags, **kws): out = Element('pandas-output') random_df = random_table() out.write(random_df.to_html(classes='pd-table')) def show_stats(*ags, **kws): out = Element('pandas-ops') input_div = document.getElementById("pandas-output") html_data = input_div.children[0].outerHTML df = pd.read_html(html_data,flavor='lxml')[0] stats_df = df.describe() out.write(stats_df.to_html(classes='pd-table')) def download_analyze(*ags, **kws): url=document.getElementById('file-url').value try: df = pd.read_csv(open_url(url)) pyscript.write('url-file-head',df.head(5).to_html(classes='pd-table')) pyscript.write('url-file-ops',df.describe().to_html(classes='pd-table')) except: pyscript.write('url-file-head',"Could not access the file")