Collect measurements from a batch¶
Bring saved measurements from many images into one table, then export them to CSV, TSV or Excel without reprocessing the images. Continuing in a VIPP results workflow is optional.
New in 0.16.0a1
Collection combines and exports saved measurements. It does not calculate plots or statistical tests; use connected results nodes for descriptive summaries and figures.
Collect and review¶
- Finish a batch run that saves a measurement table through Batch Output. Keep its run records beside the outputs.
- In Batch workflow → Run & results, choose Collect measurement results… and select the table output to collect. For example, collect the object measurements across all images, rather than mixing them with skeleton branches or a summary table.
- Review the item list. VIPP checks that saved files match their recorded contents and that their columns, value types and units are compatible. Read the reported reasons for any unusable result.
- Add missing condition, sample or replicate annotations where useful. Select several items to apply the same annotation together. Existing annotations are retained; collection does not change the measurement values.
- Review any exclusions explicitly, then choose Export results… and select CSV, TSV or Excel. You do not need to save a VIPP collection or open another workflow first. The review stays open after export.
A failed or missing result is not an image with zero objects. A verified table with no rows remains in the collection's item inventory, but contributes no invented object row. Changed, missing, unsupported, failed and excluded results remain distinguishable, so a smaller collected table cannot silently look like a complete experiment.
Older runs may need to be repeated
Older CSV/TSV outputs may lack the recorded value types needed for safe
collection. VIPP will explain this rather than guessing whether, for
example, 001 is a sample name or the number 1. To collect those results,
deliberately rerun the analysis with a VIPP build that records the required
information. Opening the collection review never reruns images for you.
Choose an export¶
- CSV or TSV: one measurement table for other analysis software. The
optional image summary companion is selected by default; keep it to
preserve each image's outcome, annotations and inclusion decision, including
images with no measurement rows. It uses the same filename with
-image-summarybefore the extension, for examplenuclei-image-summary.csvbesidenuclei.csv. - Excel (
.xlsx): one workbook with Measurements, Image summary and About this collection sheets. The image summary is always present; About this collection explains the measurement units and records run information.
The image summary is an inventory, not a new statistical analysis. An empty valid result and an excluded or failed image remain different outcomes. If you omit the CSV/TSV companion, the measurement rows alone cannot show all images that contributed no rows.
CSV/TSV does not store VIPP's column types or units, and spreadsheet apps may
reinterpret identifiers such as 001. Check import settings before analysing
those files. The Excel workbook keeps text as text and records units, but
spreadsheet number precision still applies. Save the native VIPP collection
as well when exact values and their original types must be retained.
Optionally continue in VIPP¶
Choose Save VIPP collection… separately to keep a reusable
.vipp-results.json dataset, such as
nuclei-measurements.vipp-results.json. It preserves VIPP's typed measurements,
units and complete collection records for reopening; CSV/TSV and Excel are
exports for use elsewhere, not replacements for this native dataset.
Open VIPP collection in a results workflow is unchecked by default. Select it before saving only if you want a new VIPP results workflow. Saving without it does not open another workflow, and the collection review stays open.
Table Source provides the saved measurement rows to the ordinary table tools:
- Select Table Columns keeps the fields you need, including identity keys.
- Add Metadata Columns adds further constant annotations.
- Summarize Measurements calculates its existing grouped summaries.
In the Table Source inspector, Review collection… shows the saved item inventory read-only, including empty results and exclusions. Choose measurement dataset… lets you select a saved dataset to use in that node. Its ordinary Export table… action exports just the node's table as CSV or TSV. For the Excel workbook or the image-summary companion, use Export results… in the batch collection review instead.
Collection appends observations from different images. Merge Tables has a different job: joining columns that describe matching observations. Neither equal row counts nor repeated local object numbers prove a match. Keep the source/image identifiers so object 1 in one image is not confused with object 1 in another.
Choose sample and replicate names to reflect the experiment—for example, an animal, culture or independent experiment. Several images from one sample do not automatically become independent replicates. Adding these labels does not itself establish which statistical comparisons are appropriate.
Save, move or share¶
The .vipp-results.json file contains measurements, units, reviewed
annotations and an inventory of included and omitted results with their run
records. It contains no input images, masks, mesh geometry or previews.
Review source names and annotations before sharing; they can contain private
information.
Save the results workflow separately. Its JSON references the dataset's path and recorded hash; it does not embed the table. Keep the dataset with the workflow when moving or sharing them. VIPP reports a missing or changed dataset instead of silently accepting different measurements. Opening a collection does not require recalculating the source images.
A reproducibility package still excludes result files. Share this measurement dataset separately if you want the recipient to inspect it. For column identities and units, see Measurement tables and units.
Python export is not yet available for Table Source workflows. Use VIPP to reopen and calculate the saved results workflow. A reproducibility package explains this limitation and omits the Python runner for these workflows; ordinary image-workflow exports are unchanged.