Results

Data visualization panel with plots, charts, images, and tables

Overview #

The Results panel provides a comprehensive data visualization API for creating plots, charts, images, and tables. Results are sent from Python via WebSocket and displayed in real-time with version control, tagging, and interactive viewing features.

Quick Start #

from witwin import Results
import numpy as np
 
# Create a simple line plot
x = np.linspace(0, 10, 100)
y = np.sin(x)
Results.plot(x, y, title="Sine Wave")
 
# Commit to save the results
Results.commit(message="My first plot")

API Reference #

Basic Plot Types #

from witwin import Results
 
# Line plot
Results.plot(x, y, title=None, xlabel=None, ylabel=None)
 
# Bar chart
Results.bar(categories, values, title=None)
 
# Scatter plot
Results.scatter(x, y, title=None)
 
# Heatmap/Image
Results.imshow(data, title=None, cmap=None)
 
# Data table
Results.table(data_dict, title=None)

Statistical Plot Types #

# Histogram
Results.hist(data, bins=30, title=None)
 
# Pie chart
Results.pie(values, labels, title=None)
 
# Box plot
Results.boxplot(data, title=None)

Advanced Plot Types #

# Area chart
Results.area(x, y, title=None)
 
# Radar chart
Results.radar(values, labels, title=None)
 
# 3D Surface plot
Results.surface(x, y, z, title=None)
 
# Contour plot
Results.contour(x, y, z, title=None)
 
# Error bar plot
Results.errorbar(x, y, yerr=None, xerr=None, title=None)
 
# Stem plot
Results.stem(x, y, title=None)
 
# Polar plot
Results.polar(theta, r, title=None)
 
# Text display
Results.text(text, title=None)

Utility Methods #

# Clear pending plots
Results.clear()
 
# Commit results to version history
Results.commit(message=None, tag=None, tag_color=None)
 
# Delete a specific version
Results.delete_version(version_id)
 
# Get all version history
Results.get_all_versions()

Plot Examples #

Line Plot #

from witwin import Results
import numpy as np
 
x = np.linspace(0, 10, 100)
y = np.sin(x)
Results.plot(x, y, title="Sine Wave", xlabel="Time", ylabel="Amplitude")
Results.commit()

Bar Chart #

categories = ['Model A', 'Model B', 'Model C', 'Model D']
values = [0.945, 0.923, 0.887, 0.912]
Results.bar(categories, values, title="Model Accuracy Comparison")
Results.commit()

Scatter Plot #

x = np.random.randn(100)
y = 2 * x + np.random.randn(100) * 0.5
Results.scatter(x, y, title="Linear Correlation")
Results.commit()

Heatmap/Image #

# 2D data as heatmap
data = np.random.randn(20, 30)
Results.imshow(data, title="2D Field Distribution")
 
# RGB image (H x W x 3)
image = np.random.rand(100, 100, 3) * 255
Results.imshow(image.astype(np.uint8), title="RGB Image")
Results.commit()

Data Table #

table_data = {
    'Model': ['ResNet', 'VGG', 'MobileNet'],
    'Accuracy': [0.945, 0.923, 0.887],
    'Latency (ms)': [32, 28, 45]
}
Results.table(table_data, title="Model Comparison")
Results.commit()

Histogram #

data = np.random.randn(1000)
Results.hist(data, bins=50, title="Normal Distribution")
Results.commit()

Pie Chart #

values = [35, 25, 20, 20]
labels = ['Category A', 'Category B', 'Category C', 'Category D']
Results.pie(values, labels, title="Distribution")
Results.commit()

Box Plot #

data = [np.random.randn(100) + i for i in range(4)]
Results.boxplot(data, title="Distribution Comparison")
Results.commit()

Radar Chart #

values = [0.9, 0.7, 0.8, 0.6, 0.85]
labels = ['Accuracy', 'Speed', 'Memory', 'Latency', 'Throughput']
Results.radar(values, labels, title="Model Performance")
Results.commit()

Error Bar Plot #

x = np.arange(5)
y = [2.3, 4.5, 3.2, 5.1, 4.8]
yerr = [0.3, 0.5, 0.2, 0.4, 0.3]
Results.errorbar(x, y, yerr=yerr, title="Measurements with Error")
Results.commit()

Polar Plot #

theta = np.linspace(0, 2 * np.pi, 100)
r = 1 + np.cos(theta)
Results.polar(theta, r, title="Polar Pattern")
Results.commit()

Version Management #

Commit System #

# Simple commit (creates auto-incrementing version number)
Results.commit()
 
# Commit with message
Results.commit(message="Baseline Results")
 
# Commit with tag and custom color
Results.commit(
    message="Best Model - Epoch 50",
    tag="best",
    tag_color="#00ff00"
)

Version History Features #

Persistent: Survives page refresh and reconnection
Ordering: Reverse chronological (newest first)
Deletable: Delete versions with confirmation
Tag Colors: Auto-generated or custom hex colors

Tag System #

Tags provide visual categorization with automatic or custom colors:

Tag	Use Case
`train`	Training results
`val`	Validation results
`test`	Test results
`baseline`	Baseline comparison
`best`	Best performing model
`exp`	Experimental results

# Auto color from tag name
Results.commit(tag="train")
 
# Custom color
Results.commit(tag="best", tag_color="#00ff00")

Data Type Support #

The Results API automatically converts common data types:

Input Type	Conversion
PyTorch Tensor	→ Python list
NumPy Array	→ Python list
Pandas DataFrame	→ Dict format
Python list/tuple	Direct use

Example with Different Types #

import torch
import numpy as np
 
# PyTorch tensors (auto-converted)
x = torch.linspace(0, 10, 100)
y = torch.sin(x)
Results.plot(x, y, title="PyTorch Data")
 
# NumPy arrays (auto-converted)
x = np.linspace(0, 10, 100)
y = np.sin(x)
Results.plot(x, y, title="NumPy Data")
 
Results.commit()

Usage Workflows #

Machine Learning Training #

from witwin import Results
 
for epoch in range(epochs):
    train_loss = train_one_epoch()
    val_loss = validate()
 
    # Clear previous plots
    Results.clear()
 
    # Plot training curves
    Results.plot(train_losses, title="Training Loss")
    Results.plot(val_losses, title="Validation Loss")
 
    # Show metrics
    Results.bar(['Accuracy', 'Precision', 'Recall', 'F1'],
                [acc, prec, recall, f1], title="Metrics")
 
    # Commit with epoch tag
    Results.commit(
        message=f"Epoch {epoch+1}/{epochs}",
        tag="train"
    )

Data Analysis #

from witwin import Results
 
# Distribution analysis
Results.scatter(data.x, data.y, title="Data Distribution")
Results.hist(data.values, title="Value Histogram")
 
# Correlation heatmap
Results.imshow(correlation_matrix, title="Correlation Matrix")
 
# Summary statistics
Results.table(stats_dict, title="Summary Statistics")
 
Results.commit(message="Data Analysis", tag="analysis")

Comparing Models #

from witwin import Results
 
for model_name, model in models.items():
    predictions = model.predict(test_data)
    metrics = evaluate(predictions, test_labels)
 
    Results.clear()
    Results.bar(list(metrics.keys()), list(metrics.values()),
                title=f"{model_name} Performance")
    Results.commit(message=model_name, tag=model_name.lower())

Frontend Features #

Panel UI #

Version Sidebar: List of all committed versions with tags
Plot Area: Flexible grid layout for multiple plots
Size Slider: Adjust plot size (50%-200%)
Fullscreen Mode: Expand panel to full window

Performance Guidelines #

Limit	Recommended
Plots per version	~50 plots
Data points per plot	~10,000 points
Version history	Unlimited