streamlit-swipecards 0.2.6

Streamlit Swipecards allow you to add interactive swipeable cards to your app. Supports both image cards and table row swiping with cell highlighting.

Streamlit Swipe Cards

A modern card-swiping component for Streamlit. Build swipe interfaces using images or rows from a data table.

Features

• Stacked card interface with smooth animations
• Works on both touch and mouse devices
• Like, pass and undo actions
• Optional table view powered by AG‑Grid
• Cell, row and column highlighting support
• Automatic table caching for improved performance when loading the same dataset multiple times

Installation

pip install streamlit-swipecards

Quick demo

Launch the sample app to see both modes in action:

streamlit run example.py

Usage

Image cards

import streamlit as st
from streamlit_swipecards import streamlit_swipecards

cards = [
    {"name": "Alice", "description": "Loves hiking", "image": "https://.../alice.jpg"\},
    {"name": "Bob", "description": "Chef and foodie", "image": "https://.../bob.jpg"\},
]

result = streamlit_swipecards(cards=cards, display_mode="cards", key="people")
if result:
    st.json(result)

Table cards

Provide a list of card dictionaries. Each card points to the dataset and the row it should display. Highlighting and centering options are set per card.

csv_path = "sample_data.csv"

cards = [
    {
        "dataset_path": csv_path,
        "row_index": 0,
        "name": "Alice Johnson",
        "description": "Engineering professional from New York",
        "highlight_cells": [{"row": 0, "column": "Salary", "color": "#FFB6C1"}],
        "center_table_row": 0,
        "center_table_column": "Salary",
    },
    {
        "dataset_path": csv_path,
        "row_index": 1,
        "name": "Bob Smith",
        "description": "Sales professional from California",
        "highlight_cells": [{"row": 1, "column": "Rating", "color": "#98FB98"}],
        "center_table_row": 1,
        "center_table_column": "Rating",
    },
]

result = streamlit_swipecards(cards=cards, display_mode="table", key="table")
if result:
    st.json(result)

Card dictionaries

Image card

{
    "name": "Alice",       # required
    "description": "Text", # required
    "image": "URL or base64"
}

Table card

{
    "dataset_path": "data.csv",               # required
    "row_index": 0,                            # row to display
    "name": "Row title",                      # optional
    "description": "Row description",         # optional
    "highlight_cells": [{"row": 0, "column": "Salary", "color": "#FFB6C1"}],
    "highlight_rows": [{"row": 0, "color": "#E3F2FD"}],
    "highlight_columns": [{"column": "Rating", "color": "#E8F5E8"}],
    "center_table_row": 0,
    "center_table_column": "Salary"
}

API reference

streamlit_swipecards(
    cards=None,
    dataset_path=None,         # legacy single-dataset mode
    highlight_cells=None,
    highlight_rows=None,
    highlight_columns=None,
    display_mode="cards",      # "cards" or "table"
    center_table_row=None,
    center_table_column=None,
    key=None,
) 

Return value

The component returns a dictionary:

{
    "swipedCards": [{"index": 0, "action": "right"}, ...],
    "lastAction": {"cardIndex": 0, "action": "right"},
    "totalSwiped": 3,
    "remainingCards": 7
}

How to Use

1. Swipe right 💚 or click the like button to like a card
2. Swipe left ❌ or click the pass button to pass on a card
3. Click back ↶ to undo your last action
4. Cards stack behind each other for a realistic experience
5. Smooth animations provide visual feedback

Performance

The component automatically caches loaded datasets to improve performance when multiple cards reference the same file. This means:

• No redundant file reads: The same dataset file is only loaded once, even if multiple cards use it
• Automatic cache invalidation: When a file is modified, the cache is automatically updated
• Memory efficient: Uses Streamlit's built-in @st.cache_data for optimal memory management

To see the performance benefits in action, run:

python demo_caching.py

---

Released under the MIT License.