chore: one licence, a citation file, synthetic-data notes and a portable benchmark

.github/CONTRIBUTING.md

@@ -224,8 +224,9 @@ The Python is tooling, not a shipped package, and it is not covered by CI:
   what a change did to real output.
 - `benchmarks/make_tutorial_dataset.py` generates the dataset used by the
   tutorial and by the example report.
-- `benchmarks/amap_performance_benchmark.py` profiles a run on a full local
-  dataset; `benchmarks/README.md` documents its invocation and its output.
+- `benchmarks/performance_benchmark.py` times a run on a pangenome you have
+  (wall time, peak memory, disk); `benchmarks/README.md` documents its
+  invocation and its output.
 - `benchmarks/recore_report.py` and `tools/make_logo_variants.py` generate
   assets.
 
@@ -282,14 +283,9 @@ By contributing you agree that your contribution is licensed under the same
 terms as the project, which are the ones in [LICENSE](../LICENSE) at the
 repository root.
 
-> [!WARNING]
-> The repository currently disagrees with itself about which licence that is.
-> `Cargo.toml`, the README and the documentation all say GNU General Public
-> License v3.0 (`GPL-3.0-only`), while the committed `LICENSE` file contains the
-> text of the GNU **Affero** General Public License v3. Those are different
-> licences with different obligations for anyone running a modified copy over a
-> network. This needs resolving by the maintainer, one way or the other, before
-> the first outside contribution lands.
+That licence is the GNU Affero General Public License v3.0 (`AGPL-3.0-only`).
+`Cargo.toml`, `CITATION.cff`, the README and the website name it too, and
+`tools/check_site.py` fails if one of them drifts from `LICENSE`.
 
 Participation is covered by the [Code of Conduct](CODE_OF_CONDUCT.md).
 Security issues do not go in the issue tracker: see [SECURITY.md](SECURITY.md).

.gitignore

@@ -25,10 +25,6 @@ tutorial_out/
 out_strict/
 out_balanced/
 out_sensitive/
-benchmarks/amap_performance_metrics.tsv
+benchmarks/performance_metrics.tsv
 
 
-# Added by cargo
-
-/target
-.claude/

CHANGELOG.md

@@ -88,6 +88,25 @@ and Recore uses semantic versioning.
 
 ### Changed
 
+- Recore is licensed under the GNU Affero General Public License v3.0
+  (`AGPL-3.0-only`), the licence its `LICENSE` file holds; `Cargo.toml`, the
+  README and the website said GPL-3.0. A `CITATION.cff` names the authors for
+  GitHub's citation button. `tools/check_site.py` fails if `Cargo.toml`,
+  `CITATION.cff`, the README, the contributing guide or the website name a
+  licence other than the one in `LICENSE`.
+- The example report says it is made from synthetic data: next to its own
+  title, where it shows in the website's frame, full screen and on a phone,
+  and next to every link to it on the website, in the README and in the
+  documentation. `tools/label_example_report.py` adds the note after the report
+  is regenerated, and `tools/check_site.py` fails on a report or a link without
+  it.
+- `benchmarks/performance_benchmark.py` times Recore on any pangenome, given
+  the same inputs as Recore, with an optional tree to keep a subset of genomes
+  and a cluster limit for a quick run. It replaces a benchmark that only ran on
+  one local dataset. `benchmarks/self_benchmark.py` works in the system's
+  temporary directory instead of a macOS path.
+- The Run report page no longer prints a leftover MkDocs button attribute as
+  text, and `tools/check_site.py` fails on MkDocs syntax left in the markdown.
 - The documentation site is now a short website instead of fourteen MkDocs
   pages: an install line, a command builder that writes the run for your
   pangenome tool and the files you have, a one-minute example with the counts it

CITATION.cff

@@ -0,0 +1,23 @@
+cff-version: 1.2.0
+message: "If you use this software, please cite it as below."
+title: "Recore"
+abstract: "Post-hoc rescue and quality control of pangenome presence/absence matrices."
+type: software
+authors:
+  - family-names: Ruiz-Rodriguez
+    given-names: Paula
+    orcid: "https://orcid.org/0000-0003-0727-5974"
+    affiliation: "I2SysBio, University of Valencia-CSIC, FISABIO Joint Research Unit Infection and Public Health, Valencia, Spain"
+    email: paula.ruiz.rodriguez@csic.es
+  - family-names: Coscolla
+    given-names: Mireia
+    orcid: "https://orcid.org/0000-0003-0752-0538"
+    affiliation: "I2SysBio, University of Valencia-CSIC, FISABIO Joint Research Unit Infection and Public Health, Valencia, Spain"
+repository-code: "https://github.com/PathoGenOmics-Lab/recore"
+url: "https://pathogenomics-lab.github.io/recore/"
+license: AGPL-3.0-only
+keywords:
+  - pangenome
+  - presence/absence matrix
+  - Panaroo
+  - bioinformatics

Cargo.toml

@@ -5,7 +5,7 @@ edition = "2021"
 rust-version = "1.85"
 authors = ["Paula Ruiz-Rodriguez <paula.ruiz.rodriguez@csic.es>", "Mireia Coscolla"]
 description = "Post-hoc rescue of falsely absent genes in pangenome matrices"
-license = "GPL-3.0-only"
+license = "AGPL-3.0-only"
 repository = "https://github.com/PathoGenOmics-Lab/recore"
 keywords = ["pangenome", "panaroo", "core-genome", "rescue", "bioinformatics"]
 categories = ["science", "command-line-utilities"]

README.md

@@ -21,7 +21,7 @@ puts back the genes that are really there, and keeps the evidence for each one.
 <sub>For Panaroo, Roary, PPanGGOLiN, PanTA and PIRATE · pure Rust · Linux and macOS</sub>
 
 <a href="https://github.com/PathoGenOmics-Lab/recore/actions/workflows/ci.yml"><img src="https://github.com/PathoGenOmics-Lab/recore/actions/workflows/ci.yml/badge.svg" alt="CI"></a>
-<a href="LICENSE"><img src="https://img.shields.io/badge/license-GPLv3-blue.svg" alt="License: GPLv3"></a>
+<a href="LICENSE"><img src="https://img.shields.io/badge/license-AGPLv3-blue.svg" alt="License: AGPLv3"></a>
 <img src="https://img.shields.io/badge/Rust-1.85%2B-orange.svg" alt="Rust 1.85+">
 
 </div>
@@ -105,8 +105,8 @@ Four files answer most questions. [All of them](docs/guide/outputs.md).
 | `orthology_evidence.tsv` | One row per change: the evidence, the decision, and wording you can use in a paper. |
 | `recore_report.html` | The whole run in one page that works offline. Open it in any browser. |
 
-[Explore a real report](https://pathogenomics-lab.github.io/recore/#report),
-the one the example run writes.
+[Explore an example report](https://pathogenomics-lab.github.io/recore/#report),
+the one the example run writes from synthetic data.
 
 ## 5 · One setting that matters
 
@@ -146,7 +146,7 @@ here as well:
 > Ruiz-Rodriguez P, Coscolla M. (2026). Recore: post-hoc rescue and quality
 > control of pangenome presence/absence matrices. *In preparation.*
 
-<sub>GPL-3.0, see [LICENSE](LICENSE) · [PathoGenOmics Lab](https://github.com/PathoGenOmics-Lab) · Paula Ruiz-Rodriguez, paula.ruiz.rodriguez@csic.es</sub>
+<sub>AGPL-3.0, see [LICENSE](LICENSE) · [PathoGenOmics Lab](https://github.com/PathoGenOmics-Lab) · Paula Ruiz-Rodriguez, paula.ruiz.rodriguez@csic.es</sub>
 
 ---
 

benchmarks/README.md

@@ -155,43 +155,38 @@ cross-references, a product and a long note), six names per CDS, which is what
 an annotation index has to hold for most real assemblies; that run goes to
 `benchmarks/search_scaling_rich_metrics.tsv`.
 
-## AMAP performance benchmark
+## Performance on your own pangenome
 
-`amap_performance_benchmark.py` profiles Recore on the full local AMAP
-pangenome. It is a resource benchmark, not a precision/recall truth set.
+`performance_benchmark.py` times Recore on a pangenome you have: wall time,
+peak memory of the whole process tree, temporary disk, output size, and the
+rescue counts, appended to `benchmarks/performance_metrics.tsv`. It is a
+resource benchmark, not a truth set.
 
-The script:
-
-- filters `AMAP-gene_presence_absence.csv` with the canonical 459-tip tree
-- links FASTA and GFF inputs instead of copying them
-- runs the release binary with the publication settings
-- records wall time, peak RSS, temporary disk use, output size, rescues, and
-  corrected matrix entries
-
-Run from the repository root:
+Give it the same inputs as Recore. A Newick tree keeps only the genomes that are
+its tips, and `--max-clusters` keeps the first clusters, to time a subset; both
+need a Panaroo or Roary `gene_presence_absence.csv`. FASTA and GFF files are
+linked, not copied. The `publication` profile adds `--synteny`,
+`--multi-allele` and `--nscluster`; `rescue_only` searches without them.
 
 ```bash
-python3 benchmarks/amap_performance_benchmark.py \
-  --threads 8 \
-  --profile publication \
-  --minimap2 /path/to/minimap2
+python3 benchmarks/performance_benchmark.py \
+  -p gene_presence_absence.csv -r pan_genome_reference.fa \
+  -g genomes/ -a annotations/ --tree species.nwk --threads 8
 ```
 
-For a quick smoke test:
+For a quick smoke test on the tutorial dataset (below):
 
 ```bash
-python3 benchmarks/amap_performance_benchmark.py \
-  --max-clusters 50 \
-  --threads 4 \
-  --profile rescue_only \
-  --minimap2 /path/to/minimap2 \
-  --metrics /tmp/recore_amap_smoke_metrics.tsv
+python3 benchmarks/performance_benchmark.py \
+  -p tutorial_data/gene_presence_absence.csv \
+  -r tutorial_data/pan_genome_reference.fa \
+  -g tutorial_data/genomes --max-clusters 20 --profile rescue_only \
+  --threads 4 --metrics recore_smoke_metrics.tsv
 ```
 
-The default AMAP output metrics file is
-`benchmarks/amap_performance_metrics.tsv`. Because it records local hardware
-and temporary paths, treat it as a local run log rather than a portable
-benchmark table.
+It builds `target/release/recore` first unless you pass `--recore` with a
+binary or `--skip-build`. The metrics file records the machine and temporary
+paths, so treat it as a local run log rather than a portable benchmark table.
 
 ## Tutorial dataset