Replace DataFrames's default _repr_html_ (closes #76)

src/arche/__init__.py

@@ -5,11 +5,14 @@
 SH_URL = "https://app.scrapinghub.com/p"  # noqa
 
 from arche.arche import Arche
+from arche.tools import dataframe
 from arche.tools.schema import basic_json_schema
 import numpy as np
 import pandas as pd
 import plotly.io as pio
 
+pd.DataFrame._repr_html_ = dataframe._repr_html_
+
 pio.renderers.default = "notebook_connected+jupyterlab"
 
 __all__ = ["basic_json_schema", "Arche", "np", "pd"]

src/arche/tools/dataframe.py

@@ -0,0 +1,13 @@
+import re
+
+
+def make_urls_clickable(val):
+    if isinstance(val, str) and re.search("^https?://", val):
+        return f'<a target="_blank" href="{val}">{val}</a>'
+    else:
+        return val
+
+
+def _repr_html_(self, *args, **kwargs):
+    styler = self.style.format(make_urls_clickable)
+    return styler.render(*args, **kwargs)

tests/tools/test_dataframe_html_output.py

@@ -0,0 +1,26 @@
+import pandas as pd
+import pytest
+
+
+@pytest.fixture()
+def df_with_urls():
+    data = {"col1": [1, 2], "col2": ["http://foo.com", "https://bar.com"]}
+    return pd.DataFrame(data)
+
+
+def test_df_has_clickable_urls(df_with_urls):
+    html = df_with_urls._repr_html_()
+    assert '<a target="_blank" href="http://foo.com">http://foo.com</a>' in html
+    assert '<a target="_blank" href="https://bar.com">https://bar.com</a>' in html
+
+
+def test_derivaded_df_has_clickable_urls(df_with_urls):
+    html = df_with_urls.head()._repr_html_()
+    assert '<a target="_blank" href="http://foo.com">http://foo.com</a>' in html
+    assert '<a target="_blank" href="https://bar.com">https://bar.com</a>' in html
+
+
+def test_arche_df_does_not_add_links_if_no_url_found():
+    df = pd.DataFrame({"col1": [1, 2], "col2": ["foo", "bar"]})
+    html = df._repr_html_()
+    assert "<a href=" not in html