See your Pywr model as the network it is

Pywr for VS Code turns a long JSON file into a schematic you can draw on, forms you can fill in, a run you can follow and charts you can check, all in the editor. It works with Pywr v2 and Pywr v1.

Install it free Read the guide

Free, for VS Code 1.110 or later. Already in VS Code? Install it from there.

Pywr for VS Code: the sidebar lists the nodes and parameters of a model, and the schematic shows the results of a run on the network, with the volume of the reservoir and the flow of each node.

Draw the network on a schematic

Open the schematic and every node and edge of your model is on the page. What you change on it is written to the model file.

Drag nodes, connect them to add edges and drop new ones from the toolbar
Nodes with coordinates are drawn on a map
Show the results of a run on the network, one time step at a time

Draw the schematic

The network of a model drawn on a map in VS Code: a catchment feeds a reservoir, which supplies a town through a treatment works, next to a river with an intake.

Edit a component in a form

Click a component to edit all its fields in one place. Saving the form writes the change back into the JSON file.

Pick names from lists and files with a file picker
Profiles, control curves and bathymetry are charted as you type
CSV tables and timeseries are previewed with their columns

Edit with forms

Catch mistakes while you type

The model is checked against the Pywr schema as you edit it, and for the mistakes a schema cannot see.

A type that does not exist or a missing field is underlined at once
So is a name nothing defines, a node with no edges or a list of the wrong length
The light bulb offers the closest name, or creates the component

Check the model for mistakes

A misspelt parameter name underlined in the JSON editor, with a quick fix menu offering to replace it with the closest name or to create the parameter.

Run the model with one button

Press Run and follow its progress, or run it live on the schematic and watch each time step. Each run saves its results in a new dated folder next to the model.

Pywr v2 is built in: you only need a Python environment
Pick the solver, the scenarios and the output folder without editing the model
See how long each part of the model took

Run the model

A live run on the schematic: a bar above the network steps through the dry scenario, at day 5 of 731, and each node shows its flow or volume for that day.

Chart the results where you ran them

Open an HDF5 or CSV results file and tick what to plot, without Python or a notebook.

Time series, duration curves and percentile bands across scenarios
Aggregate by month or year, and read the statistics of each line
Compare two runs, then export the chart as PNG or the values as CSV

Chart the results

The results viewer charting the volume of a reservoir over two years for a dry, an average and a wet scenario.

Check the model still behaves

A QA report is a set of charts you define once and draw again after every change, such as a reservoir against its control curve.

Runs a copy of the model and exports only what is on your charts
Saved in a file next to the model, to commit and share with it
Exports the charts as a PDF document

Build a QA report

A QA report with three charts: the storage of a reservoir with its drought curve, the supply to a town against its demand, and the flow of a river at a gauge.

Two versions of Pywr, one extension

The extension recognises which version a model is written for and shows it in the status bar. The sidebar, the forms, the schematic and the charts work the same way for both.

	Pywr v2	Pywr v1
Recognised by	The `metadata`, `timestepper` and `network` sections	A `timestepper` section and a list of `nodes`
Checked against	The Pywr schema, and rules the schema cannot express	Its own schema of the pywr types, and the rules pywr applies when it loads a model
Runs with	The Pywr engine built into the extension	pywr 1 installed in your Python environment
Solvers	CLP, HiGHS, CBC and microlp	GLPK, GLPK edge and lp_solve
Results charted	HDF5 and CSV outputs	Files of the `TablesRecorder` and the `CSVRecorder`
Schematic	Positions saved in the model, as a diagram or on a map	The positions, icons and edge colours saved in the model are kept

Whether you read JSON or not

If you would rather not open the file

You can build and maintain a model from the sidebar, the forms and the schematic, and never write a brace.

A form for every node, parameter, table, timeseries and output
Search any component by name or by type
See what a component uses, and what uses it, before you change it
Rename or delete, and every reference follows

If you live in the JSON

The editor knows the components of your model, like it knows the symbols of code.

F2 renames a component everywhere it is used
F12 goes to a component, Shift F12 lists its usages
Names and types complete as you type, and hovering explains them
Links above each component open its form or show it on the schematic

The rest of VS Code works as usual around the model. You can commit it and compare its versions with Git, write its Python scripts with the Python extension, and use the terminal, the search across files and any other extension you have installed.

Install it, then open a model

Install Pywr for VS Code, free, from the Extensions view of VS Code or with the button below.
Open the folder with your model, then its JSON file. No model yet? Run Pywr: New model and pick a template.
Open the schematic from the editor toolbar, or press Run.

Install it free Build your first model

Needs VS Code 1.110 or later. To run a model you also need a Python environment with Python 3.10 or later, and pywr 1 installed in it for Pywr v1 models. Charting results needs no Python.