Quickstart

Build a batch

The code below will create a mesh sensitivity batch and highlight some of the key features. The comments in the code should help to clarify the process.

import OrcFxAPI as ofx  # why so much typing?
import numpy as np  # need some numpy magic
from qalx_orcaflex.core import QalxOrcaFlex, OrcaFlexBatch, ModelSource  # core classes
from qalx_orcaflex import data_models as dm  # get data_models with little typing

qfx = QalxOrcaFlex()  # initialise a pyqalx session for OrcaFlex

m = ofx.Model()  # empty model
line = m.CreateObject(ofx.otLine, "My Line")  # make a line
line.Length[0] = 200  # make the line long
line.EndAX = 0  # move end A
line.EndBX = 150  # move end B
line.EndAY, line.EndBY, line.EndAZ, line.EndBZ = (0, 0, 0, 0)  # put everything else at 0
line.TargetSegmentLength[0] = 0.5  # initial segment length

batch_options = dm.BatchOptions(
    batch_queue="example batch bot",  # queue our batch bot is looking at for jobs
    sim_queue="example sim bot",  # queue our sim bot is looking at for jobs
)

ofx_batch = OrcaFlexBatch(
    "Example sensitivity batch",  # name of the batch
    session=qfx,  # qalx session we created earlier
    batch_options=batch_options,  # options for this batch
    meta={"example": "mesh_sensitivity"},  # some metadata (could be anything here)
    verbose=True,  # show some print statements so we can see what's happening
)

results = [  # I want some results
    dm.RangeGraph(  # a range graph
        object="My Line",  # of line
        variable="Curvature",  # curvature
        period=dm.Period(named_period="Latest Wave")  # for the latest wave
    ),
    dm.RangeGraph(  # and a range graph
        object="My Line",  # for line
        variable="Effective tension",  # tension
        period=dm.Period(named_period="Latest Wave")  # for the latest wave
    ),

]

with ofx_batch as batch:  # let's build this batch
    for l_seg in np.linspace(0.01, 0.5, 10):  # I want 10 segment lengths equally spaced between 0.01 and 0.5
        line.TargetSegmentLength[0] = l_seg  # set the segment length
        case_info = dm.RawInfo(key="Segment Length", value=l_seg)  # define some load case information
        batch.add(
            source=ModelSource(m, f"seg@{l_seg:2.2}"),  # add the model with a name that shows the segmentation
            load_case_info=dm.LoadCaseInfo(raw_info=[case_info]),  # add the load case info
            required_results=results  # add the results
        )

# Once the code has exited the above `with` block then the batch has been built and submitted

Start the bots

We use the qalx command line tool to start the BatchBot and SimBot

qalx bot-start --queue-name "example batch bot" qalx_orcaflex.bots:batch_bot

In a separate command line window

qalx bot-start --queue-name "example sim bot" qalx_orcaflex.bots:sim_bot

Read the results

Once your jobs have finished you can get the results summary and references to all the jobs and results:

from qalx_orcaflex.core import QalxOrcaFlex

qfx = QalxOrcaFlex()

mesh_batch_results = qfx.get_batch_results_summary("Example sensitivity batch")

for result_name, range_graph in mesh_batch_results['Range Graphs'].items():
    print(f"{result_name}: Max {range_graph['max_value']:2.2f} @ {range_graph['arc_max_value']:2.2f}")
    worst_case_info = range_graph['max_case_info']
    print(f"worst segment length={worst_case_info['Segment Length']}m")

Where’s a nice pretty user interface?

Short answer, there isn’t one. There are a few reasons for this:

• User interfaces are hard to get right and we didn’t want to just ship a poor one

• We expect a lot of users to already have their own interfaces and systems that they would integrate with

• Custom Interfaces shows how easy it is to create your own