Batch Processing¶
Batch processing lets you apply the same pipeline to hundreds of files automatically.
How It Works¶
- Folder Iterator provides one file path per iteration
- A reader node loads each file (Image Reader, Table Reader, etc.)
- Your processing pipeline runs on each file
- Batch Accumulator collects results across all iterations
- Downstream nodes (tables, plots) run once after all files are processed
Setting Up a Batch Pipeline¶
Step 1: Add a Folder Iterator¶
- Add a Folder Iterator node
- Set the Folder to your data directory
- Set the Pattern to match your files (e.g.,
*.tif,*.csv,*.png)
Step 2: Build Your Pipeline¶
Connect the Folder Iterator to a reader node, then to your processing nodes as usual.
Step 3: Collect Results¶
Add a Batch Accumulator node between your pipeline output and any display/plot nodes.
The Batch Accumulator:
- Collects one table per iteration during the batch
- Merges all tables into one combined table after the batch completes
- Adds a
filecolumn with the source filename
Step 4: Run¶
Click Batch Run (Ctrl+B). Progress is shown in the status bar.
Example Pipelines¶
Image measurement¶
Measure particle properties across a folder of microscopy images.
Folder Iterator → Image Reader → Gaussian Filter → Threshold → Particle Props → Batch Accumulator → Data Table
CSV aggregation¶
Load multiple CSV files, filter rows, and combine into a single summary table.
Folder Iterator → Table Reader → Filter Table → Aggregate Table → Batch Accumulator → Data Table
Multi-file statistics¶
Run the same statistical test on each dataset and collect all p-values.
Folder Iterator → Table Reader → Group Comparison → Batch Accumulator → Sort Table → Data Table
Batch figure export¶
Generate and save a plot for each input file.
Folder Iterator → Table Reader → Scatter Plot → Data Saver
Batch Gate¶
The Batch Gate node pauses the batch pipeline at each iteration so you can review results and adjust node settings before continuing. Wire it anywhere in your pipeline as a checkpoint.
- Next — let the current iteration continue past the gate
- Refresh — re-evaluate upstream nodes (useful after changing a parameter)
- Pass All — stop pausing and let the remaining iterations run automatically
Tips¶
Skipping the batch loop
If you've already run a batch and want to re-process only the downstream analysis (e.g., change a plot), just click Run Graph (Ctrl+W). Synapse detects that the accumulators already have data and skips the per-file loop.
Stopping a batch
Click the Stop button to cancel. The batch stops after the current file finishes processing.
Memory
For very large batches (1000+ files), ensure you have sufficient RAM. Each iteration's results are held in memory until the batch completes.