This repository hosts a dataset for interfacial electrochemistry data, focusing on cyclic voltammograms of well-defined single crystal electrodes recorded in aqueous and non-aqueous electrolytes. It consists of published data that is either retraced from source PDFs or provided by the authors, and can be browsed on echemdb.org/cv.
The data is provided as frictionless-based unitpackages
and originates from two kinds of input:
- Digitized data retraced from figures in publications, stored as SVG and
YAML files and converted with
svgdigitizer. - Directly submitted data provided by authors as CSV files together with YAML metadata.
In both cases a BibTeX (BIB) reference ties the data to its publication, and all input YAML files and output Data Packages are validated against the echemdb-metadata schema. The dataset is generated automatically on each release from the input files in this repository and is maintained by the echemdb community.
The generated data archive can be downloaded directly as a ZIP:
- GitHub release (latest): data-0.9.0.zip
- Zenodo (archived, citable): download the ZIP from the Files section of the Zenodo record, which always resolves to the latest version.
All releases are also listed in the release section.
A collection can be created from the the echemdb module of the unitpackages interface
(see unitpackages installation instructions).
from unitpackage.database.echemdb import Echemdb
db = Echemdb.from_remote()Install the latest version of the module.
pip install git+https://github.com/echemdb/electrochemistry-data.gitIn your preferred Python environment retrieve the URL with the data via
from echemdb_ecdata.url import ECHEMDB_DATABASE_URL
ECHEMDB_DATABASE_URLYou can contribute data in two ways:
- Digitize data from a publication. The preparation of the files and the
extraction of the data from a PDF source is described in the
svgdigitizer workflow. The resulting
SVG and YAML files go into
literature/svgdigitizer/. - Submit your own raw data directly. Add your measurement as a CSV file
together with a YAML metadata file under
literature/source_data/; no digitization is required.
If you want to work on the data and repository itself, install pixi and clone the repository:
git clone https://github.com/echemdb/electrochemistry-data.git
cd electrochemistry-dataFor possible commands run
pixi runMore pixi tasks can be inferred from the pyproject.toml.
The repository converts source data into standardized frictionless datapackages:
# Convert all data (SVG digitizer + raw data)
pixi run -e dev convert
# Convert only SVG digitizer data (from literature/svgdigitizer/)
pixi run -e dev convert-svg
# Convert only raw data (from literature/source_data/)
pixi run -e dev convert-raw
# Clean generated data before converting
pixi run -e dev clean-dataA typical workflow:
# Clean previous builds and convert all data
pixi run -e dev clean-data && pixi run -e dev convertGenerated datapackages are written to data/generated/svgdigitizer/ and data/generated/source_data/.
Both SVG digitization and raw data conversion use a batch approach that imports heavy dependencies once and processes all files in a single Python process. This avoids the ~3 s Python startup overhead per file that occurs when spawning a subprocess for each file, reducing full-rebuild time from ~15 min to ~30-50 s for 273 SVG files.
Force a full rebuild (ignoring timestamps):
pixi run -e dev convert-forceVerify that the batch conversion produces output identical to existing generated data:
pixi run -e dev verify-svg # SVG digitizer output
pixi run -e dev verify-raw # Source data output
pixi run -e dev verify-all # Both at onceAll data (input YAML and output JSON) is validated against the echemdb-metadata schema. In addition, filenames, identifiers, and bibliography keys are validated for consistency.
Two umbrella tasks cover all checks:
# Validate all input files (YAML schema, filenames/identifiers, bib keys)
pixi run -e dev validate-input
# Validate all generated files (JSON schema, identifiers)
pixi run -e dev validate-generatedThese are also used in the CI workflows. You can run individual sub-tasks:
# Schema validation
pixi run -e dev validate-svgdigitizer-yaml # Input YAML (svgdigitizer)
pixi run -e dev validate-source-yaml # Input YAML (source data)
pixi run -e dev validate-svgdigitizer # Generated JSON (svgdigitizer)
pixi run -e dev validate-raw # Generated JSON (source data)
# Filename and identifier validation
pixi run -e dev validate-identifiers # All input filenames
pixi run -e dev validate-svgdigitizer-filenames # SVG digitizer filenames only
pixi run -e dev validate-source-filenames # Source data filenames only
pixi run -e dev validate-generated-identifiers # Generated data identifiers
# Bibliography key validation
pixi run -e dev validate-bib-keys # Check bib keys match expected identifiers
pixi run -e dev validate-bib-utf8 # Check for LaTeX accent encodingsValidate against a specific schema version:
pixi run -e dev validate-input --version 0.8.0
pixi run -e dev validate-generated --version main# Lowercase SVG labels and filenames (enforced for Windows compatibility)
pixi run -e dev fix-lowercase # Apply changes
pixi run -e dev fix-lowercase-dry-run # Preview only
# Convert LaTeX accent encodings to UTF-8 in bibliography.bib
pixi run -e dev fix-bib-utf8 # Apply changes
pixi run -e dev fix-bib-utf8-dry-run # Preview only
# Auto-fix identifier mismatches (detects dir name != YAML citationKey)
pixi run -e dev fix-identifiers # Apply changes
pixi run -e dev fix-identifiers-dry-run # Preview only
# Rename directories and files after a bib key change (manual)
pixi run -e dev rename-identifiers OLD_NAME NEW_NAME
# Migrate input YAML metadata across breaking metadata-schema releases
# (applies the migration steps shipped with the metadata-schema repository)
pixi run -e dev migrate-metadata # Apply changes
pixi run -e dev migrate-metadata-dry-run # Preview onlyNew literature and source-data pull requests can be reviewed with the help of an AI assistant. Both assistants run the same checks (input validation, the review module, and PDF cross-checks against the cited paper) and produce a review report.
- GitHub Copilot — invoked as slash commands:
/review-source-data-pr <folder-name>/review-literature-pr <folder-name>
- Claude Code — the equivalent skills:
review-source-data-pr <folder-name>review-literature-pr <folder-name>
Here <folder-name> is the entry directory (e.g. droog_1980_oxygen_387). The
review generates a REVIEW.md in the repository root (the parent directory of the
entries). This file is git-ignored and must not be committed: the reviewer marks
each finding as accept / reject / comment and adds notes directly in it, and the
assistant then applies the accepted fixes.
For recurring issues indicate by (add to context) in the comment, that these aspects should be considered in future reviews.
This will update the review skills.
The canonical, shared instructions for both assistants live in
.github/prompts/ (the Claude skills under
.claude/skills/ delegate to the same prompt files).
You can redistribute and/or modify the contents of this repository under the terms of the GNU General Public License (GPL) as published by the Free Software Foundation; either version 3.0 of the License, or (at your option) any later version (GPL-3.0-or-later). See https://www.gnu.org/licenses.
The data and generated data, i.e., the contents of
literature/ and the contents of the generated data archive
that is published with each release, are also available under either of the
following two licenses, at your option:
- The Creative Commons Attribution 4.0 International License (CC-BY-4.0)
- The Open Data Commons Attribution License (ODC-By-1.0)
While you should consult the above licenses for all the technical details, the above roughly translates to the following in practice:
- You may use any of the raw and processed data you find here as long as you give credit.
- If you want to incorporate any source code you find here into your published product or project it must itself be released under the GPL version 3 or later.