v0.1.2 - Add DataFrame support

Author: cedricvlt

README.md

@@ -3,7 +3,7 @@ Based on [react-awesome-query-builder](https://github.com/ukrbublik/react-awesom
 
 Check out [live demo](https://condition-tree-demo.streamlit.app/) !
 
-This component allows users to build complex condition trees that can be used, for example, to filter a dataframe or build a query.
+This component allows users to build complex condition trees that can be used to filter a dataframe or build a query.
 
 <img src="preview.jpg" width="500" alt="preview">
 
@@ -33,11 +33,37 @@ This component allows users to build complex condition trees that can be used, f
 
 ## Basic usage
 
+### Filter a dataframe
+
+```python
+import pandas as pd
+from streamlit_condition_tree import condition_tree, config_from_dataframe
+
+# Initial dataframe
+df = pd.DataFrame({
+    'First Name': ['Georges', 'Alfred'],
+    'Age': [45, 98],
+    'Favorite Color': ['Green', 'Red'],
+    'Like Tomatoes': [True, False]
+})
+
+# Basic field configuration from dataframe
+config = config_from_dataframe(df)
+
+# Condition tree
+query_string = condition_tree(config)
+
+# Filtered dataframe
+df = df.query(query_string)
+```
+
+### Build a query
+
 ```python
 import streamlit as st
 from streamlit_condition_tree import condition_tree
 
-
+# Build a custom configuration
 config = {
     'fields': {
         'name': {
@@ -58,11 +84,13 @@ config = {
     }
 }
 
+# Condition tree
 return_val = condition_tree(
     config,
     return_type='sql'
 )
 
+# Generated SQL
 st.write(return_val)
 ```
 
@@ -72,18 +100,22 @@ st.write(return_val)
 
 ```python
 def condition_tree(
-    config: Dict
-    return_type: str
-    tree: Dict
-    min_height: int
-    placeholder: str
+    config: dict,
+    return_type: str,
+    tree: dict,
+    min_height: int,
+    placeholder: str,
     key: str
 )
 ```
 
-- **config**: Python dictionary that resembles the JSON counterpart of
-  the React component [config](https://github.com/ukrbublik/react-awesome-query-builder/blob/master/CONFIG.adoc).  
-*Note*: Javascript functions (ex: validators) are not yet supported.
+- **config**: Python dictionary (mostly used to define the fields) that resembles the JSON counterpart of
+  the React component.
+   
+A basic configuration can be built from a DataFrame with `config_from_dataframe`.  
+For a more advanced configuration, see the component [doc](https://github.com/ukrbublik/react-awesome-query-builder/blob/master/CONFIG.adoc)
+and [demo](https://ukrbublik.github.io/react-awesome-query-builder/).  
+  *Note*: Javascript functions (ex: validators) are not yet supported.
 
 
 - **return_type**: Format of the returned value :
@@ -92,9 +124,9 @@ def condition_tree(
   - sql
   - spel
   - elasticSearch
-  - jsonLogic
-    
-  Default : queryString
+  - jsonLogic  
+
+  Default : queryString (can be used to filter a pandas DataFrame using DataFrame.query)
 
 
 - **tree**: Input condition tree (see section below)
@@ -125,5 +157,4 @@ It can be loaded as an input tree through the `tree` parameter.
 
 
 ## Potential future improvements
-- **Dataframe filtering support**: automatically build config from dataframe and return a query string adapted to `pandas.DataFrame.query`
 - **Javascript support**: allow injection of javascript code in the configuration (e.g. validators)

setup.py

@@ -7,7 +7,7 @@
 
 setuptools.setup(
     name="streamlit-condition-tree",
-    version="0.1.1",
+    version="0.1.2",
     author="Cédric Villette",
     author_email="[email protected]",
     description="Condition Tree Builder for Streamlit",

streamlit_condition_tree/__init__.py

@@ -15,6 +15,43 @@
     _component_func = components.declare_component("streamlit_condition_tree", path=build_dir)
 
 
+type_mapper = {
+    'b': 'boolean',
+    'i': 'number',
+    'u': 'number',
+    'f': 'number',
+    'c': '',
+    'm': '',
+    'M': 'datetime',
+    'O': 'text',
+    'S': 'text',
+    'U': 'text',
+    'V': ''
+}
+
+
+def config_from_dataframe(dataframe):
+    """Return a basic configuration from dataframe columns"""
+
+    fields = {}
+    for col_name, col_dtype in zip(dataframe.columns, dataframe.dtypes):
+        col_type = 'select' if col_dtype == 'category' else type_mapper[col_dtype.kind]
+
+        if col_type:
+            col_config = {
+                'label': col_name,
+                'type': col_type
+            }
+            if col_type == 'select':
+                categories = dataframe[col_name].cat.categories
+                col_config['fieldSettings'] = {
+                    'listValues': [{'value': c, 'title': c} for c in categories]
+                }
+            fields[f'{col_name}'] = col_config
+
+    return {'fields': fields}
+
+
 def condition_tree(config: dict,
                    return_type: str = 'queryString',
                    tree: dict = None,
@@ -32,12 +69,16 @@ def condition_tree(config: dict,
         Format in which output should be returned to streamlit.
         Possible values : queryString | mongodb | sql | spel |
         elasticSearch | jsonLogic.
+        Default : queryString (compatible with DataFrame.query)
     tree: dict or None
         Input condition tree
+        Default: None
     min_height: int
         Minimum height of the component frame
+        Default: 400
     placeholder: str
         Text displayed when the condition tree is empty
+        Default: empty
     key: str or None
         An optional key that uniquely identifies this component. If this is
         None, and the component's arguments are changed, the component will
@@ -51,6 +92,16 @@ def condition_tree(config: dict,
 
     """
 
+    if return_type == 'queryString':
+        # Add backticks to fields with space in their name
+        fields = {}
+        for field_name, field_config in config['fields'].items():
+            if ' ' in field_name:
+                field_name = f'`{field_name}`'
+            fields[field_name] = field_config
+
+        config['fields'] = fields
+
     output_tree, component_value = _component_func(
         config=config,
         return_type=return_type,
@@ -60,6 +111,11 @@ def condition_tree(config: dict,
         placeholder=placeholder,
         default=['', '']
     )
+
+    if return_type == 'queryString' and not component_value:
+        # Default string that returns all the values in DataFrame.query
+        component_value = 'index in index'
+
     st.session_state[key] = output_tree
 
     return component_value

streamlit_condition_tree/example.py

@@ -1,30 +1,26 @@
+import numpy as np
+import pandas as pd
 import streamlit as st
-from streamlit_condition_tree import condition_tree
+from streamlit_condition_tree import condition_tree, config_from_dataframe
 
+df = pd.read_csv(
+    'https://media.githubusercontent.com/media/datablist/sample-csv-files/main/files/people/people-100.csv',
+    index_col=0,
+    parse_dates=['Date of birth'],
+    date_format='%Y-%m-%d')
+df['Age'] = ((pd.Timestamp.today() - df['Date of birth']).dt.days / 365).astype(int)
+df['Sex'] = pd.Categorical(df['Sex'])
+df['Likes tomatoes'] = np.random.randint(2, size=df.shape[0]).astype(bool)
 
-config = {
-    'fields': {
-        'name': {
-            'label': 'Name',
-            'type': 'text',
-        },
-        'qty': {
-            'label': 'Age',
-            'type': 'number',
-            'fieldSettings': {
-                'min': 0
-            },
-        },
-        'like_tomatoes': {
-            'label': 'Likes tomatoes',
-            'type': 'boolean',
-        }
-    }
-}
+st.dataframe(df)
+
+config = config_from_dataframe(df)
 
 return_val = condition_tree(
     config,
-    return_type='sql'
 )
 
-st.write(return_val)
+st.code(return_val)
+
+df = df.query(return_val)
+st.dataframe(df)

streamlit_condition_tree/frontend/package.json

@@ -1,6 +1,6 @@
 {
   "name": "streamlit-condition-tree",
-  "version": "0.1.1",
+  "version": "0.1.2",
   "private": true,
   "dependencies": {
     "@fontsource/source-sans-pro": "^5.0.8",