Examples¶
The following examples show how to use the loader APIs:
1. Parsing¶
This sample shows how to read a CWL document from a remote public URL:
In [1]:
Copied!
import cwl_loader
cwl_document = cwl_loader.load_cwl_from_location('https://raw.githubusercontent.com/eoap/application-package-patterns/refs/heads/main/cwl-workflow/pattern-1.cwl')
import cwl_loader
cwl_document = cwl_loader.load_cwl_from_location('https://raw.githubusercontent.com/eoap/application-package-patterns/refs/heads/main/cwl-workflow/pattern-1.cwl')
2025-08-14 23:17:23.210 | INFO | cwl_loader:load_cwl_from_location:192 - Loading CWL document from https://raw.githubusercontent.com/eoap/application-package-patterns/refs/heads/main/cwl-workflow/pattern-1.cwl...
2025-08-14 23:17:23.289 | INFO | cwl_loader:load_cwl_from_yaml:123 - Updating the model to v1.2...
2025-08-14 23:17:23.293 | INFO | cwl_loader:load_cwl_from_yaml:134 - Raw CWL document successfully updated to v1.2! Now converting to the CWL model...
2025-08-14 23:17:23.358 | INFO | cwl_loader:load_cwl_from_yaml:142 - Raw CWL document successfully updated to v1.2! Now dereferencing the FQNs...
2025-08-14 23:17:23.359 | INFO | cwl_loader:_clean_process:65 - Cleaning Workflow pattern-1...
2025-08-14 23:17:23.360 | INFO | cwl_loader:_clean_process:65 - Cleaning CommandLineTool clt...
2025-08-14 23:17:23.360 | INFO | cwl_loader:load_cwl_from_yaml:150 - CWL document successfully dereferenced!
1.1 Check the parsed document¶
In [2]:
Copied!
import sys
cwl_loader.dump_cwl(cwl_document, sys.stderr)
import sys
cwl_loader.dump_cwl(cwl_document, sys.stderr)
cwlVersion: v1.2
$graph:
- id: pattern-1
class: Workflow
label: Water bodies detection based on NDWI and the otsu threshold
doc: Water bodies detection based on NDWI and otsu threshold applied to a
single Landsat-8/9 acquisition
inputs:
- id: aoi
label: area of interest
doc: area of interest as a bounding box
default: -118.985,38.432,-118.183,38.938
type: string
- id: epsg
label: EPSG code
doc: EPSG code
default: EPSG:4326
type: string
- id: bands
label: bands used for the NDWI
doc: bands used for the NDWI
default:
- green
- nir08
type:
name: _:cd7c7d82-bbc9-4519-8be7-3ef62c0ba226
items: string
type: array
- id: item
label: Landsat-8/9 acquisition reference
doc: Landsat-8/9 acquisition reference
type: Directory
outputs:
- id: water_bodies
label: Water bodies detected
doc: Water bodies detected based on the NDWI and otsu threshold
outputSource:
- step/stac-catalog
type: Directory
requirements: []
cwlVersion: v1.2
steps:
- id: step
label: Water bodies detection
doc: Water bodies detection based on NDWI and otsu threshold applied to a
single Landsat-8/9 acquisition
in:
- id: item
source: item
- id: aoi
source: aoi
- id: epsg
source: epsg
- id: band
source: bands
out:
- stac-catalog
run: '#clt'
$namespaces: &id001
s: https://schema.org/
- id: clt
class: CommandLineTool
inputs:
- id: item
type: Directory
inputBinding:
prefix: --input-item
- id: aoi
type: string
inputBinding:
prefix: --aoi
- id: epsg
type: string
inputBinding:
prefix: --epsg
- id: band
type:
- name: _:399f9633-e10f-4638-bec2-ee5841478319
items: string
type: array
inputBinding:
prefix: --band
outputs:
- id: stac-catalog
type: Directory
outputBinding:
glob: .
requirements:
- class: InlineJavascriptRequirement
- class: EnvVarRequirement
envDef:
- envName: PATH
envValue: /app/envs/runner/bin
- class: ResourceRequirement
coresMax: 1
ramMax: 512
hints:
- class: DockerRequirement
dockerPull: docker.io/library/runner:latest
cwlVersion: v1.2
baseCommand:
- runner
arguments:
- pattern-1
$namespaces: *id001
2. Components diagram¶
This sample shows how to write a CWL document to a stream (string, file, ...):
In [3]:
Copied!
from cwl2puml import (
to_puml,
DiagramType
)
from io import (
StringIO,
BytesIO
)
from IPython.display import display
from plantuml import deflate_and_encode
from PIL import Image
from urllib.request import urlopen
out = StringIO()
to_puml(
cwl_document=cwl_document,
diagram_type=DiagramType.COMPONENTS,
output_stream=out
)
clear_output = out.getvalue()
encoded = deflate_and_encode(clear_output)
diagram_url = f"https://www.plantuml.com/plantuml/png/{encoded}"
with urlopen(diagram_url) as url:
img = Image.open(BytesIO(url.read()))
display(img)
from cwl2puml import (
to_puml,
DiagramType
)
from io import (
StringIO,
BytesIO
)
from IPython.display import display
from plantuml import deflate_and_encode
from PIL import Image
from urllib.request import urlopen
out = StringIO()
to_puml(
cwl_document=cwl_document,
diagram_type=DiagramType.COMPONENTS,
output_stream=out
)
clear_output = out.getvalue()
encoded = deflate_and_encode(clear_output)
diagram_url = f"https://www.plantuml.com/plantuml/png/{encoded}"
with urlopen(diagram_url) as url:
img = Image.open(BytesIO(url.read()))
display(img)
2.1 Check the generated PlantUML code¶
In [4]:
Copied!
print(clear_output)
print(clear_output)
@startuml
skinparam linetype ortho
node "Workflow 'pattern-1'" {
component "pattern-1" as pattern_1 {
portin "aoi" as pattern_1_aoi
portin "epsg" as pattern_1_epsg
portin "bands" as pattern_1_bands
portin "item" as pattern_1_item
portout "water_bodies" as pattern_1_water_bodies
}
component "step" as pattern_1_step {
portin "item" as pattern_1_step_item
pattern_1_item .down.> pattern_1_step_item
portin "aoi" as pattern_1_step_aoi
pattern_1_aoi .down.> pattern_1_step_aoi
portin "epsg" as pattern_1_step_epsg
pattern_1_epsg .down.> pattern_1_step_epsg
portin "band" as pattern_1_step_band
pattern_1_bands .down.> pattern_1_step_band
portout "stac-catalog" as pattern_1_step_stac_catalog
}
}
node "CommandLineTool 'clt'" {
component "clt" as clt {
portin "item" as clt_item
portin "aoi" as clt_aoi
portin "epsg" as clt_epsg
portin "band" as clt_band
portout "stac-catalog" as clt_stac_catalog
}
}
pattern_1_step_stac_catalog .up.> pattern_1_water_bodies
pattern_1_step .right.> clt
@enduml