Reference
Files BIDS Manager writes
What a project folder holds after a conversion, file by file: the dataset itself, and the .bidsmgr folder where BIDS Manager keeps its history, scan tables, logs and reports.
The dataset
| File or folder | Written by | What it is |
|---|---|---|
dataset_description.json |
creating the project; the conversion; the metadata step | The dataset's name, BIDS version, authors, licence, and what generated it. |
README, CHANGES |
creating the project (README); the metadata step (both) | A starting README to replace with a description of the dataset, and a change log. |
.bidsignore |
creating the project | Tells validators to ignore .bidsmgr/ and the conversion's scratch folder. |
participants.tsv, participants.json |
the metadata step | One row per subject, and the description of each column. |
sub-<label>/ |
the conversion | The recordings, each with its sidecar and companion files. |
sub-<label>/ |
the metadata step | Each recording of the session with its acquisition time. |
task-<label>_bold.json and other sidecars at the top |
Tools > Fix ups... > Move shared fields up | Sidecars shared by many files (inheritance). |
phenotype/ |
the metadata step, with a phenotype spreadsheet | Questionnaires and measures, one table per instrument. |
CITATION.cff |
Tools > Fix ups...; bidsmgr-metadata --write-citation; the post-conversion setting |
How to cite the dataset, from its description. Optional. |
sourcedata/ |
the conversion, when Keep in sourcedata/ is on | The raw files of each subject, as they were or as one archive per series, with a manifest.tsv of checksums. |
derivatives/bidsmgr-qc/ |
the quality check | Its results, per image, so they are not computed twice. |
derivatives/ |
Tools > Remove the skull... | Skull-stripped images, as a derivative dataset. |
The .bidsmgr folder
BIDS Manager's own folder in the dataset. It is listed in .bidsignore, so validators skip it, and it is what makes scans resumable and every change traceable. Leave it out when you share the dataset (Sharing a dataset).
.bidsmgr/
├── project/
│ ├── meta.json the project: name, BIDS Manager and BIDS versions, creation time
│ ├── events.jsonl what happened to the project, one event per line
│ └── scans/
│ └── 0001__raw/ one scan table
│ ├── inventory.tsv the scan table, as last saved
│ ├── inventory.tsv.recording_meta.json the dataset metadata you gave
│ ├── events.jsonl every edit made to this table, for Undo
│ ├── version.json where the raw data was, and the table's state
│ └── ... the scan's caches (series files, features)
├── editor/
│ ├── operations.log every change made in the Editor, one per line
│ └── originals/ the files as they were before each change
├── logs/ one log per scan, conversion, metadata step and validation
├── provenance/ per subject: what the conversion did, with which tools
├── errors/ per subject, when its conversion failed: the reason
├── backup/ files an update conversion replaced
├── merged/<source>/ the history of a subject merged into another
├── adopted_manifest.json for an adopted dataset: every file as it was found
├── accepted_findings.json validation findings you accepted
├── viz/scenes/ viewer scenes you saved
├── metadata_report.json what the metadata step filled, and every TODO it wrote
├── validation_report.json the last validation's findings
└── validation_report.html the same, as a page to read or send
During a conversion each subject is written into .tmp_bidsmgr/sub-<label>/ in the dataset first, and moved into place only when it converted completely. The folder is removed afterwards; if you find one, a conversion was interrupted, and it can be deleted.
Outside the dataset
| What | Where |
|---|---|
| Your settings and recent projects | your user account (First launch) |
| Logs of commands run outside a project | ~/.bidsmgr/logs/ |
| A scan without a project | the scan table file you name, with its caches beside it (inventory.tsv.files_ and similar) |
BIDS Manager writes nothing into the raw folder you scan.