Skip to content

Importing Data

FlowGraph imports external data through reader nodes. Add the reader that matches your source. Every mesh reader has a required path parameter of type Text, exported as an input port. A file can be selected with a Select file node and connected to that port, or the path can be entered directly in a workflow.

Files can also be added to the project from File → Add files.... The native file picker accepts multiple selections. FlowGraph then asks whether to copy the files into a temporary project folder or link to the original locations. Copies are removed when the application session ends; linked originals are never deleted by FlowGraph.

Mesh readers return a mesh output containing FlowGraph's canonical Mesh document. The CSV reader returns a table output containing a mapping of column names to NumPy arrays.

Choosing a reader

Reader node Use it for Input Output Dependency
Read Mesh (Muscat) Mesh formats supported by Muscat's reader factory path parameter/input (Text) mesh (Mesh document) Muscat
Read Mesh (MeshIO) Mesh formats supported by meshio path parameter/input (Text) mesh (Mesh document) meshio and Muscat's meshio bridge
Read Mesh (MeshLane) Mesh formats supported by MeshLane path parameter/input (Text) mesh (Mesh document) MeshLane and the Muscat MeshLane bridge
Read Table (Pandas) Comma-separated tabular data path (Text) table (Table) pandas

Choose Read Mesh (Muscat) when the format is supported by Muscat directly and you want Muscat's extension-aware reader selection. Choose Read Mesh (MeshIO) or Read Mesh (MeshLane) when the format is better supported by one of those libraries. Use Read Table (Pandas) for CSV data rather than a mesh reader.

Read Mesh (Muscat)

The Muscat reader uses the file extension to select a Muscat reader, then loads the file into the canonical MeshDocument used by FlowGraph. The result is the editable mesh model; visualization is built later as a derived 3D projection.

Select file.path ──▶ Read Mesh (Muscat).path
                     Read Mesh (Muscat).mesh ──▶ mesh operations

Muscat must recognize the file extension and the file contents must be valid for the selected reader. The path must point to an existing file and must have an extension.

Read Mesh (MeshIO)

The MeshIO reader calls meshio.read(path) and converts the resulting mesh through Muscat's MeshIO bridge. The output is still a FlowGraph MeshDocument, so downstream mesh nodes can use it in the same way as output from the Muscat reader.

Use this reader when meshio provides the format support you need. The meshio package and the corresponding Muscat bridge must be installed in the active environment.

Read Mesh (MeshLane)

The MeshLane reader calls MeshLane to read the source and converts the result through Muscat's MeshLane bridge. It produces the same mesh / Mesh document contract as the other mesh readers.

MeshLane is optional. If it is not installed, or if the installed Muscat version does not provide the MeshLane bridge, execution stops with an actionable error. Install FlowGraph with the meshlane extra when that extra is available in your environment.

Read Table (Pandas): CSV

The Pandas reader is intended for .csv files containing a header row. It reads the file with pandas.read_csv, then converts each column into a NumPy array. The node returns a dictionary keyed by the original column names:

table = {
    "x": numpy array,
    "temperature": numpy array,
}

The output type is Table, not Mesh document. Connect it to a node that expects tabular data; it cannot be connected directly to a mesh operation unless an explicit conversion or tabular-processing node is provided.

The CSV file must exist and have a filename extension. Ensure pandas is installed before executing the node. CSV parsing options are currently those provided by the adapter's default pandas.read_csv call.

File requirements and errors

All four readers validate the path before attempting to parse it:

  • The path must identify an existing regular file.
  • The path must include a file extension.
  • The file must be readable and valid for the selected reader.

If validation, dependency loading, or parsing fails, the node reports a reader- specific error. Check that the selected node matches the data type and format, that the file path is correct, and that optional dependencies are installed.

Keep source data in a location that remains available for the workflow.

See the interface, workflow node reference, or jump to visualization.