Node Overview

Nodes provide template structure for performing data extraction, processing, and API communication. The diagram below shows how a Node performs a base function, but exposes an interface for user-defined code to enable tailoring.

In Jupyter notebooks, Nodes that are editable by users are marked with light blue backgrounds for user-defined SQL and Python cells. These cells can be committed to a repository and deployed to a workflow orchestrator by clicking the button in the toolbar or by pressing Cmd+S (macOS) or Ctrl+S (Windows/Linux).

Key Characteristics of Nodes

Understanding how Nodes interact with files, APIs, and their input/output types is essential for effective usage. Nodes may also have a user-editable component, which is indicated in the table below.

List of Available Nodes and Key Characteristics

The table below provides a list of available Nodes, along with details on whether they include a user-editable component and if they can handle multiple inputs.

• Is Editable: Indicates that the Node includes a user-editable component.
• Is Multi: Indicates if the Node can accept multiple inputs.

Category	Name	Input Types	Output Types	Is Editable	Is Multi
Analysis	Branch_Python	Table(s) and/or File(s)	NodeReturn	True	False
Analysis	Python	Table(s) and/or File(s)	NodeReturn	True	False
Analysis	Transform_SQL	Table(s)	Table	True	False
Analysis	Trigger_Python	Table(s) and/or File(s)	FlowInputs	True	False
App	APINode	API	NodeReturn	True	False
App	AirtableExport	Table	API	True	False
App	AirtableImport	API	NodeReturn	False	False
App	Azure_Query	Table	NodeReturn	True	False
App	Azure_Read	API	FileAny	False	False
App	Azure_Read_Multi	API	FileAny	False	True
App	Azure_Write	FileAny	API	False	False
App	Benchling_Api	API	NodeReturn	True	False
App	Benchling_Event	Event	FlowInputs or NodeReturn	True	False
App	Benchling_Read	API	NodeReturn	True	False
App	Benchling_Read_Object	API	NodeReturn	True	False
App	Benchling_Warehouse_Query	Table	NodeReturn	True	False
App	Benchling_Warehouse_Sync	API	NodeReturn	True	False
App	Benchling_Write	Table	NodeReturn	True	False
App	Benchling_Write_Object	Table(s) and/or File(s)	NodeReturn	True	False
App	Coda_Write	Table	NodeReturn	True	False
App	ELabNext_Write	Table	NodeReturn	True	False
App	Load_Parquet_to_Table	API	FileAny	False	False
App	S3_Event	Event	FlowInputs	True	False
App	S3_Read	API	FileAny	True	False
App	S3_Write	FileAny	API	False	False
App	SciNote_API	Table	NodeReturn	True	False
App	Smartsheet_Read	API	NodeReturn	True	False
App	Snowflake_Write	Table	NodeReturn	False	False
File	AVI_Read	FileAVI	NodeReturn	True	False
File	AVI_Read_Multi	Set[FileAVI]	NodeReturn	True	True
File	CSV_Read	FileCSV	NodeReturn	True	False
File	CSV_Read_Multi	Set[FileCSV]	NodeReturn	True	True
File	CSV_Write	Table(s)	NodeReturn	True	False
File	Excel_Read	FileExcel	NodeReturn	True	False
File	Excel_Read_Multi	Set[FileExcel]	NodeReturn	True	True
File	Excel_Write	Table(s)	NodeReturn	True	False
File	FCS_Extract_Load	FileFCS	NodeReturn	True	False
File	HDF5_Read	FileHDF5	NodeReturn	True	False
File	Image_Read	FileImage	NodeReturn	True	False
File	Image_Read_Multi	Set[FileImage]	NodeReturn	True	True
File	Image_Write	Table(s)	NodeReturn	True	False
File	Input_File	FileAny	NodeReturn	True	False
File	Input_File_Multi	Set[FileAny]	NodeReturn	True	True
File	PDF_Read	FilePDF	NodeReturn	True	False
File	PDF_Read_Multi	Set[FilePDF]	NodeReturn	True	True
File	Powerpoint_Write	Table(s)	NodeReturn	True	False
File	XML_Read	FileXML	NodeReturn	True	False
File	Zip_Read	FileZip	NodeReturn	True	False
Instrument	Instron_Tensile_Read	FileIsTens	NodeReturn	True	False
Instrument	LCMS_Read	File	NodeReturn	True	False
Instrument	LCMS_Read_Multi	File	NodeReturn	True	True
Instrument	LC_Read	File	NodeReturn	True	False
Instrument	LC_Read_Multi	File	NodeReturn	True	True
Instrument	Profilometer_Read	FileHDF5	NodeReturn	True	False
Instrument	Synergy_Read	FileTxt	NodeReturn	True	False
Instrument	Synergy_Read_Multi	Set[FileTxt]	NodeReturn	True	True
Instrument	WSP_Read	FileWSP	NodeReturn	True	False
Tag	Benchling_Tag	TagBenchling	string	False	False
Tag	Input_Param		string	False	False

Node Categories

• App: Integrates with third-party APIs for data processing; often requires key exchange between the third-party service and Ganymede.
• Analysis: Performs data manipulations using Python or SQL.
• Instrument: Handles data from laboratory instruments.
• File: Conducts ETL operations on specified data types within the Ganymede cloud.
• Tag: Defines parameters at Flow runtime.

Input and Output Types for Nodes

NodeReturn Object

Many Nodes return a NodeReturn object, which contain tables and files for storage in the Ganymede data lake.

To initialize a NodeReturn object, the following parameters can be passed:

• param tables_to_upload: dict[str, pd.DataFrame] | None - Tables to be stored in Ganymede, keyed by name.
• param files_to_upload: dict[str, bytes] | None - Files to be stored in Ganymede, keyed by filename.
• param if_exists: str - String indicating whether to overwrite or append to existing tables in Ganymede data lake. Valid values are "replace", "append", or "fail"; defaults to "replace".
• param tables_measurement_units: Optional[dict[str, pd.DataFrame]] - (If provided) Specifies the measurement units for columns; keys are table names, values are pandas DataFrames with "column_name" and "unit" as columns.
• param file_location: str - Specifies the bucket location ("input" or "output"); required only if files_to_upload is not null, defaults to "output".
• param wait_for_job: Whether to wait for the write operation to complete before continuing execution; defaults to False.
• param tags: dict[str, list[dict[str, str]] | dict[str, str]] | None: Dictionary of files to tag, with keys as file names and values as a dictionary of keyword parameters for the add_file_tag function. Multiple tags can be added to a single file by passing a list of add_file_tag parameters in the dictionary.

NodeReturn Example

The contents of a NodeReturn object can be inspected in the notebook, where table headers and list of files are displayed. Below is an example of creating a NodeReturn object:

import pandas as pd

def execute():
    message = "Message to store in file"
    byte_message = bytes(message, "utf-8")

    df = pd.DataFrame({"col1": [1, 2, 3], "col2": [4, 5, 6]})

    # upload a table named 'my_table' and a file named 'my_file'
    return NodeReturn(files_to_upload={"my_file.txt": message}, tables_to_upload={"my_table": df})

execute()

This code produces the following summary of the NodeReturn object:

Docstrings and source code can be viewed by typing ?NodeReturn and ??NodeReturn respectively in a cell in the editor notebook.

Example: NodeReturn Object with Tags

The following code demonstrates the use of the tags parameter in a NodeReturn object:

import pandas as pd

def execute():
    message = "Message to store in file"
    byte_message = bytes(message, "utf-8")

    df = pd.DataFrame({"col1": [1, 2, 3], "col2": [4, 5, 6]})

    # Tags are added to the file 'my_file.txt'
    # Any parameters that can be passed into the add_file_tag function can be passed into the tags parameter
    # of the NodeReturn object.  For more information on the add_file_tag function, see the Tags page.
    # 
    # Note that the input_file_path parameter within the add_file_tag function does not need to be specified
    return NodeReturn(files_to_upload={"my_file.txt": message}, tables_to_upload={"my_table": df}, 
                      tags={"my_file.txt": [{"tag_type_id": "Experiment ID", "display_value": "EXP005"}]})

execute()

FlowInputs object

Nodes that trigger other Flows return a FlowInputs object, which specifies the inputs to the triggered Flow.

To initialize a FlowInputs object, use the following parameters from ganymede_sdk.io:

• param files: list[FlowInputFile] | None - Files to pass to the triggered Flow.
• param params: list[FlowInputParam] | None - Parameters to pass to triggered Flow.
• param tags: list[Tag] | None - Tags to pass to the triggered Flow.

FlowInputFile is a dataclass used for passing files to a Node. Attributes include:

• param node_name: str - Name of the Node within triggered Flow to pass file(s) into
• param param_name: str - Node parameter in the triggered Flow Node that specifies the string pattern that the filename must match (e.g., "csv" for the CSV_Read Node)
• param files: dict[str, bytes] - Files to pass into Node

FlowInputParam is a dataclass used to pass parameters into a Node. It has the following attributes:

• param node_name: str - Name of the Node within triggered Flow to pass parameter(s) to.
• param param_name: str - Node parameter in the triggered Flow Node that is used to specify the string pattern that the parameter must match (e.g., "param" for the Input_Param Node).
• param param_value: str - Value to pass into Node.

Tag is a dataclass used to pass Benchling Tags into a node, and is used exclusively with the Benchling_Tag Node. It has the following attributes:

• param node_name: str - Name of the node within triggered Flow to pass tag(s) into.
• param display_tag: str - Value displayed in the dropdown in the Ganymede UI. For Benchling_Tag Nodes, this is the name of the tag displayed in the dropdown in Flow View / Flow Editor.
• param run_tag: str - Underlying value of the tag. For Benchling_Tag Nodes, this is the Benchling ID associated with the value selected in the dropdown.

Other input/output types

Some other input and output types characteristic to Nodes are:

• Table: Tabular data retrieved from or passed to the Ganymede data lake via ANSI SQL queries or as Pandas DataFrames.
• API: Access through third-party APIs.
• File-related inputs/outputs: Specific file types, including:
  • FileAVI: AVI file
  • FileCSV: CSV file
  • FileExcel: Excel file (xls, xlsx, ..)
  • FileImage: Image file (png, bmp, ..)
  • FileHDF5: HDF5 file
  • FileXML: XML file
  • FileZip: Zip file
  • FileAny: generic data file, which may be unstructured
• TagBenchling: Benchling run tag
• string: String parameter

Python sets, lists, and dictionaries are denoted as Set, List, and Dict, respectively.

Optional indicates that the input or output is optional.

User-editable Nodes

User-editable Nodes present an interface for modifying and testing code that is executed by the workflow management system. These Jupyter notebooks are split into the following sections:

• Node Description: A brief description of the Node's functionality.
• Node Input Data: For Nodes that retrieve tabular data, this section specifies the SQL query used to fetch data for processing.
• User-Defined Function: The execute function in this section processes the data. The function is called during the Flow execution.

info

The execute function may call classes and functions found within the User-Defined Function cell.

• Testing Section: Cells in this section are for testing modifications to the SQL query and user-defined Python function. After making edits, save changes by clicking the button in the toolbar or selecting "Save Commit and Deploy" from the Kernel menu.

List of Available Nodes

Category	Name	Brief Description
Analysis	Branch_Python	Process data with Python and conditionally execute downstream Nodes
Analysis	Python	Process data with python
Analysis	Transform_SQL	SQL analysis Function
Analysis	Trigger_Python	Process data with Python and trigger subsequent flow
App	APINode	Generic API Access Node
App	AirtableExport	Export data from Ganymede data lake to Airtable
App	AirtableImport	Import data from Airtable into Ganymede data lake
App	Azure_Query	Query data from Azure SQL Server
App	Azure_Read	Read data from Azure Blob Storage
App	Azure_Read_Multi	Read all data from Azure Blob Storage
App	Azure_Write	Write data to Azure Blob storage
App	Benchling_Api	Read Benchling data into data lake
App	Benchling_Event	Capture events from Benchling for triggering flows or saving data
App	Benchling_Read	Read Benchling data into data lake using run tag
App	Benchling_Read_Object	Read Benchling data into data lake using object ID
App	Benchling_Warehouse_Query	Query Benchling Warehouse from Ganymede
App	Benchling_Warehouse_Sync	Sync Benchling Warehouse to Ganymede
App	Benchling_Write	Write to Benchling
App	Benchling_Write_Object	Write object to Benchling
App	Coda_Write	Write Coda tables
App	ELabNext_Write	Create and write eLabNext entry
App	Load_Parquet_to_Table	Create data lake table from parquet files
App	S3_Event	Capture events from AWS S3 for triggering flows
App	S3_Read	Ingest data into Ganymede data storage from AWS S3 storage
App	S3_Write	Write data to an S3 bucket
App	SciNote_API	Create and write SciNote entry
App	Smartsheet_Read	Read sheet from Smartsheet
App	Snowflake_Write	Sync tables in Ganymede data lake to Snowflake
File	AVI_Read	Read in contents of an AVI file to a table
File	AVI_Read_Multi	Read in contents of multiple avi files to a table
File	CSV_Read	Read in contents of a CSV file
File	CSV_Read_Multi	Read in contents of multiple CSV files
File	CSV_Write	Write table to CSV file
File	Excel_Read	Read Excel spreadsheet
File	Excel_Read_Multi	Read Excel spreadsheets
File	Excel_Write	Write Excel spreadsheet
File	FCS_Extract_Load	Load FCS file to data lake
File	HDF5_Read	Read HDF5 data
File	Image_Read	Process image data; store processed images to data store
File	Image_Read_Multi	Process image data for multiple images; store processed images to data store
File	Image_Write	Process tabular data; write an image to data lake
File	Input_File	Read data file and process in Ganymede
File	Input_File_Multi	Read data files and process in Ganymede
File	PDF_Read	Read in contents of an PDF file to a table
File	PDF_Read_Multi	Read in contents of multiple pdf files to a table
File	Powerpoint_Write	Process tabular data; write a powerpoint presentation to data lake
File	XML_Read	Read XML file into data lake
File	Zip_Read	Extract Zip file
Instrument	Instron_Tensile_Read	Load .is_tens file to data lake
Instrument	LCMS_Read	Read and process LCMS file in mzML format
Instrument	LCMS_Read_Multi	Read and process multiple LCMS files
Instrument	LC_Read	Read and process an Agilent Chemstation / MassStation HPLC file
Instrument	LC_Read_Multi	Read and process multiple Agilent Chemstation / MassStation HPLC files
Instrument	Profilometer_Read	Read Mx Profiler data file
Instrument	Synergy_Read	Load Synergy text file to data lake
Instrument	Synergy_Read_Multi	Load multiple Synergy texts file to data lake
Instrument	WSP_Read	Read FlowJo WSP file into data lake
Tag	Benchling_Tag	Read benchling tag
Tag	Input_Param	Input parameter into Flow