0
# Pipeline System
1

2
Workflow pipeline system for building multi-step interactive applications with directed graph navigation. Enables complex multi-stage data processing workflows, wizards, and guided user experiences.
3

4
## Capabilities
5

6
### Pipeline Management
7

8
Core classes for defining and managing workflow pipelines with multiple stages and navigation.
9

10
```python { .api }
11
class Pipeline:
12
    """
13
    Directed graph of workflow stages for building multi-step applications.
14
    
15
    A Pipeline represents a UI workflow of multiple linear or branching stages,
16
    where each stage returns a Panel object to render.
17
    
18
    Parameters:
19
    - stages: Dictionary mapping stage names to stage functions or objects
20
    - start: Name of the starting stage
21
    - debug: Whether to show debugging information
22
    - sizing_mode: How the pipeline should size itself
23
    - **params: Additional parameters
24
    
25
    Layout Components:
26
    - header: Pipeline header with navigation controls
27
    - title: Name of the current stage
28
    - error: Error display field
29
    - network: Network diagram of the pipeline graph
30
    - buttons: Navigation buttons and selectors
31
    - prev_button: Button to go to previous stage
32
    - prev_selector: Selector for branching previous stages
33
    - next_button: Button to go to next stage
34
    - next_selector: Selector for branching next stages
35
    """
36

37
class PipelineError:
38
    """
39
    Custom error type for displaying pipeline-specific error messages.
40
    
41
    Raised within pipeline stages to display custom error messages
42
    in the pipeline's error display area.
43
    """
44
```
45

46
## Usage Examples
47

48
### Basic Linear Pipeline
49

50
```python
51
import panel as pn
52

53
def stage1():
54
    return pn.Column(
55
        "# Stage 1: Data Input",
56
        pn.widgets.FileInput(accept='.csv'),
57
        pn.widgets.Button(name="Next", button_type="primary")
58
    )
59

60
def stage2():
61
    return pn.Column(
62
        "# Stage 2: Data Processing", 
63
        pn.pane.Markdown("Processing your data..."),
64
        pn.widgets.Button(name="Next", button_type="primary")
65
    )
66

67
def stage3():
68
    return pn.Column(
69
        "# Stage 3: Results",
70
        pn.pane.Markdown("Your results are ready!"),
71
        pn.widgets.Button(name="Download", button_type="success")
72
    )
73

74
# Create linear pipeline
75
pipeline = pn.pipeline.Pipeline({
76
    'input': stage1,
77
    'process': stage2, 
78
    'results': stage3
79
}, start='input')
80

81
pipeline.servable()
82
```
83

84
### Branching Pipeline
85

86
```python
87
import panel as pn
88

89
def analysis_choice():
90
    return pn.Column(
91
        "# Choose Analysis Type",
92
        pn.widgets.RadioButtonGroup(
93
            name="Analysis", 
94
            options=['Statistical', 'Machine Learning', 'Visualization'],
95
            value='Statistical'
96
        )
97
    )
98

99
def statistical_analysis():
100
    return pn.pane.Markdown("# Statistical Analysis Results")
101

102
def ml_analysis():
103
    return pn.pane.Markdown("# Machine Learning Results")
104

105
def visualization():
106
    return pn.pane.Markdown("# Data Visualization")
107

108
# Create branching pipeline
109
pipeline = pn.pipeline.Pipeline({
110
    'choice': (analysis_choice, ['stats', 'ml', 'viz']),
111
    'stats': statistical_analysis,
112
    'ml': ml_analysis,
113
    'viz': visualization
114
}, start='choice')
115
```
116

117
The pipeline system is essential for building complex multi-step applications, data processing workflows, configuration wizards, and any application requiring guided user experiences with branching logic.