History¶

[Unreleased]¶

• Sync conda env files with pyproject.toml: pin cellpycore ==0.2.4, sqlalchemy >= 2.0.0, and xlrd >= 2.0.1. Add xlrd>=2.0.1 to install requires (old .xls loaders). (#969)

• Zensical API pages no longer show raw Sphinx roles (:class:, :meth:, :func:); library docstrings use markdown code spans. (#967)

• Batch docs start from b = batch.load(...) (what b can do, including b.plot()) instead of a "Facade" heading. Public Batch methods have Shift-Tab docstrings. (#963)

• batch.load no longer treats a filefinder miss as success: cells with no raw files and no local .cellpy are FAILED (not an empty LOADED cell), find_files warns with the labels, and load warns pointing at batch.result.report(). (#962)

• Missing env_file is named in the UnderDefined remote-auth error (and a load-time warning) instead of only mentioning CELLPY_PASSWORD / CELLPY_KEY_FILENAME. (#961)

• Batch.drop (and drop_cells_marked_bad) remove the cell from the store, not just the cache, so plot / summaries / report work without a following update(). mark_as_bad stays a session flag. (#952)

• Collected summary y-axis titles include units (Charge Capacity (mAh/g), Coulombic Efficiency (%)). spread=True legends use custom_group_labels=, and facet rows follow the collected columns= order top to bottom. (#947)

• Collection.plot(backend="matplotlib") no longer raises TypeError: warn_once() missing 1 required positional argument. It keeps aliasing to the seaborn layout path (and now actually returns a figure - the y-label mapper crashed on the summary path). (#925)

• cycles_collector(b).plot(layout="per_cell") follows the shared cycle legend policy: more than legend_cycle_limit cycles (default 8, the same as the single-cell cycles_plot) get a colorbar instead of a legend hundreds of entries long. force_colorbar / force_legend override. (#928)

• Grouped summary collections plot the way they were collected: facets follow the columns= order instead of coming out alphabetical (derived series such as the CV split or a normalized retention curve keep their own order after the requested ones), custom_group_labels= reach the legend (integer and string group ids now match; an unlabelled group keeps its id), and the legend title of a grouped summary is Group rather than Cell. legend_title= and order_variables= still override. (#923)

• cellpy setup reports one line per real action instead of narrating itself. The parameter dump (init_filename, dst_file, not_relative, …) moved to --verbose, a dry run states each file once instead of twice in two formats, and --silent is now genuinely silent - it used to stop the questions but still print ~25 lines. A per-command --silent / --debug now also reaches the structured output, not only the plain lines. Copy pass over the remaining commands (edit, new, pull, serve, setup migrate): no more [cellpy] (cmd) prefixes, 80-column rules, apologies or RUNNING SOMETHING ELSE debug prints, and several bare print() calls that bypassed echo= (and --silent) now go through it. cellpy new --list shows template locations instead of Python tuples, cellpy pull with nothing selected and cellpy edit <unknown> are real usage errors (stderr, exit 2). (#891)

• cellpy info and cellpy info --check report rather than narrate. The check run is one line per check with a symbol, a short detail and a hint when it fails, closing with N of M checks passed - instead of === checking === banners, 80-column rules, a page of probe output per check and a failed!!!! line that shipped with a stray f. The probe narration is still there under --verbose. cellpy info --check now exits 1 when a check fails (it always exited 0), so it can be used in a script. A path setting that may be remote is now only skipped when the value actually is remote, so a broken local cellpydatadir is reported instead of excused. Output lines changed: [cellpy] version: X is now cellpy X, and [cellpy] -> <path> is now config <path>. (#891)

• executor="processes" writes .cellpy in the worker on a raw load and lazy-reopens on the parent, so batch.load(..., save_cellpy=True) no longer crashes with None.save. Cached None is not treated as loaded. (#920)

• Fix batch tqdm bars: thread children start when a worker runs (not on submit) and complete before close so Jupyter does not leave red ⅔ widgets; the overall bar reset()s its total after the journal exists so 4/25 no longer looks full. batch.load documents when to use executor and that config.batch.auto_use_file_list needs project to match the raw folder name exactly.
• batch.load / Batch.update show tqdm progress on a TTY or in Jupyter (progress=None auto, False off, True force, or a callable). Overall bar covers journal → search → cells → persist; per-cell bars cover copy / parse / save. executor="threads" draws one child bar per in-flight cell; processes keeps the overall bar only. The 3-arg on_progress callback is unchanged. (#916)
• CLI: new global --quiet / -q, --verbose and --no-color options. --quiet reports problems and the output you asked for (cellpy info still answers) and drops progress chatter; colour is otherwise automatic and honours NO_COLOR. cellpy run without a NAME, or with a NAME but no --journal / --key / --folder / --cellpy-project, is now a real usage error on stderr with exit code 2 — it used to print hand-made usage text (or a flag dump and an apology) to stdout and exit 255 or 0. cellpy convert reports a rejected --to on stderr. (#891)
• Figure export is discoverable from a collector: BatchCollector.to_image() returns image bytes and BatchCollector.save_figure(path) / Collection.save_figure(path) write one to disk. .save() still writes frame + meta.json only, and now says so and names the figure API. (#926)
• Batch.export_project(destination) writes a shareable .cellpy + journal bundle (2.x replacement for duplicate_cellpy_files). (#878)
• Run the real essential and full CI gates once for every PR targeting master and every master push; remove same-named docs-only no-op checks that could mask test results. (#876)
• Validate collected layout= / kind= / method= (raise on unknown values); layout='film' aliases kind='film' so apps no longer get a silent wrong line plot. (#874)
• spread_plot: mean traces get a real hovertemplate (group/cell, variable, cycle, mean, std); Upper/Lower Bound band traces use hoverinfo='skip'. (#875)
• Collector help is usable from Jupyter: summary_collector, cycles_collector, ica_collector and dva_collector spell out the kwargs people actually pass (columns, group_it, custom_group_labels, rate, cycles, mode, voltage_resolution, …) in their signature and a Google-style docstring, and point at SummaryOptions / CurveOptions / IcaOptions for the rest. Shift-Tab no longer shows a one-line wrapper and **overrides. (#924)
• summary_collector(b, family="fullcell_standard_gravimetric") (or y=, the summary_plot alias) builds the collection from a registered plot family, resolved against the first loaded cell's summary schema. Explicit columns= / options= still win; an unknown name raises the same ValueError that lists the known families. (#927)
• Fix scheduled CI: keep sqlalchemy-access Windows-only in conda env files, and install legacy-files (PyTables) in the scheduled pip matrix. (#885)
• Iterative fixes: document for devs how to add plots. (#892)
• arbin_sql_h5 two-stage load reads the export once: parse() caches the HDF frames and the temp copy for the following loader(), parse() honours refuse_copying, and when the harmonize prefetch succeeds the legacy row-wise datetime decode (whose only product is a raw frame that is discarded) is skipped. Prefetch failure still takes the full legacy path. (#902)
• Remote loads are cheaper: an OtherPath builds its credentialed filesystem once and reuses it for is_file / stat / copy, and from_raw no longer STATs a remote raw file it is about to copy (the copy raises if it is missing). Missing local files still raise NoDataFound. (#901)
• config.batch.auto_use_file_list is wired into journal_from_db / find_files: when enabled (default stays false) the raw-file directory is dumped once — scoped to the batch project by an exact join — and every cell is matched against that list instead of searching the tree per cell. A missing project folder raises with the joined path instead of producing an empty journal. (#900)
• Docs: batch.load(..., executor="threads") and the other measured speed knobs (auto_use_file_list, save_cellpy, cold reopen) are documented for agents; CONTRIBUTING warns about a mixed pip/conda pyarrow. (#903)
• Per-cell search_for_files searches sub-folders with rglob(..., files_only=True), so a remote search can use the single find -L listing instead of walking the tree per cell. (#899)
• v9 .cellpy writes parquet members with zstd level 3 so new files stay near the old DEFLATE size without a second zip compressor. Snappy members from #898-era files still load. (#912)
• v9 .cellpy writes its parquet members with ZIP_STORED instead of DEFLATE-ing already-compressed parquet, cutting seconds off every save. meta.json stays deflated; the file format is unchanged. (#898)
• Remote rglob(..., files_only=True) keeps the find -L listing when find exits 1 because a sibling directory was unreadable, instead of discarding it and falling back to the slow SFTP walk. (#897)
• Example notebooks use the 2.1 schema and public APIs (c.schema, potential / cycle_num / step_type, cellpy.readers.data_structures, registered summary_plot families). cellpy.get docstring examples load .cellpy files and pass nominal_capacity=. (#940)

[2.1.2] - 2026-08-09¶

Patch release — plotting and collect improvements for app builders, config and CLI correctness fixes, and safer cellpy-file writes. Additive, no breaking changes. The interim v2.1.1.post7 / v2.1.1.post8 tags and the v2.1.2a1– v2.1.2a4 pre-releases are folded into this section.

Plotting and collect¶

• PlotFamily.summary_options(hdr) returns ready SummaryOptions, so collect_summaries(batch, options=family.summary_options(hdr)) produces the columns a family declares. Families now carry the CV partition they need (*_split_constant_voltage and the full-cell families no longer depend on the caller knowing to pass partition_by_cv=True), and the synthetic mod_01_* retention column is materialised by a real transform. On the demo cell the number of registered summary families an app can actually plot goes from 8/20 (defaults) to 15/20 — the rest lack *_absolute source data. transforms() is documented as the summary_plot normalization spec, not a collect transform. (#868)
• New cellpy.collect.normalize_column_on_max, the counterpart of normalize_column for normalising against a column's own maximum. (#868)
• raw_plot can now be bounded: cycles= selects cycles (as the other plot families already allow) and max_points= thins the traces with min/max decimation per bucket, so spikes survive where plain striding would drop them. Both are plumbed through RawPrepareConfig, and the cycle filter runs before the frame is copied. On the bundled demo cell plot_type="full" drops from 18 MiB of figure JSON to 0.33 MiB at max_points=5000. (#867)
• cellpy.collect.collect_dva: multi-cell DVA collection (mirrors collect_ica), returning a Collection with the same Collection.plot() / grouping / save entry points; dva_collector convenience wrapper. (#863)
• In-memory static figure export: collection.to_image / cellpy.plotting.write_image (PNG/SVG/PDF bytes). (#818)
• ICA plotter: honour direction for line layouts, including direction='both'. (#821)
• ica_plot / dva_plot: dash discharge (dotted) vs charge (solid) when direction='both' so the two half-cycles stay distinguishable on a static export. (#862)
• Pretty-print cycles collector facet strips (Cycle N / cell label, not cycle_num=). (#820)
• Honour share_y / match_axes on collected summary spread_plot. (#817)
• Bug fix (collect): group_it=True averages multi-member groups even when some groups are singletons. (#816)

Configuration and CLI¶

• cellpy info, cellpy edit config and cellpy info --check now act on the config file that is actually loaded (cellpy.toml before a legacy .conf), and name a shadowed legacy file instead of pointing at it. (#851)
• cellpy info --configloc names a project cellpy.toml when one applies (and outranks the user file). (#853)
• config.override() is thread-/task-local via contextvars (no cross-talk between concurrent jobs). (#850)
• Security: config file dump/load no longer persist or accept legacy Arbin SQL_PWD / SQL_UID under [instruments] (env-only credentials). (#849)
• Keep default cellpy setup off the reader stack: --check / --deps are opt-in; --no-deps deprecated. (#839)
• Speed up CLI cold start: lazy package/CLI imports so cellpy info --version no longer loads the full reader stack. (#837)

Data and files¶

• Atomic .cellpy / .h5 writes: save stages next to the destination and replaces it only when complete, so an interrupted save no longer corrupts or destroys the file. (#845)
• refresh_after + SUMMARY_META_DEPENDENCIES: rebuild meta-dependent summary columns after mass / area / nominal-capacity / cycle-mode edits without a full make_summary(). (#846)
• cellpy.get: when instrument= is set, do not auto-pick the native .h5 / .hdf5 format. (#819)

Documentation¶

• Tutorial notebooks are no longer duplicated. The top-level examples/ folder holds the single maintained copy — the one the docs link to as a download — and docs/examples/ now contains only generated markdown, figure directories and screenshots. dev/render_example_notebooks.py renders from examples/ into docs/examples/, covering every tutorial, so 08_batmo_bdf and 09_loading_pec_data gain documentation pages. The batch tutorial moves to examples/batch_utility/cellpy_batch_processing.ipynb (page: examples/batch_utility/cellpy_batch_processing.md), and committed run artifacts under docs/examples/ (cellpy files, dumps, output CSVs) are gone. 09_loading_pec_data no longer sets the removed prms.Reader. (#869)

• Release-prep documentation sweep. The Incremental capacity analysis and Batch processing tutorials taught API that was removed in 2.1 and could no longer be run as written; both notebooks were rewritten onto the current API, re-executed, and re-rendered, so their code, tables and figures are real again (prms.Paths -> cellpy.config.paths, collectors.Batch*Collector -> cellpy.collect with .show() -> .plot(), ica.dqdv_cycle / dqdv_cycles / dqdv_np -> ica.dqdv, y="dq" -> y="dqdv", get_cap().voltage -> .potential, and polars-aware access to b.summaries). Also corrected the frame-type claim in The fundamentals of cellpy (pandas per-cell, polars in cellpy.collect, parquet on disk), nine factual errors in the About loaders developers-guide section, and "deprecated, removal in 2.1" wording in the migration guides for API that has since been removed. (#866)

Chores¶

• Ignore stray local leftovers cellpy_batch_test_batch.json and testdata/hdf5/20160805_test001_45_cc.cellpy. (#855)

[2.1.1.post6] - 2026-08-02¶

• Batch load: AUTO prefers existing .cellpy (new NEWEST for freshness checks), force_recalc remakes steps/summary, and skip rewriting cellpy files already loaded from disk. (#825)

[2.1.1.post4] - 2026-07-31¶

Post-release of 2.1.1 — batch JSON load path, metadata peek, ingestion-form schema.

• cellpy.instrument_meta_schema(instrument) describes cellpy.get metadata knobs for building per-instrument ingestion forms. (#800)

• Lightweight cellpy.read_meta(path) peeks v9 / HDF5 cellpy-file metadata without loading raw/steps/summary frames. (#799)

• Batch: load BatBase / custom JSON journals via cellpy.batch.load(..., db_reader=...) with file search after read (reader= alias). (#345)

• Add issue-flow as a uv dev dependency so agents get the CLI from uv sync / uv run issue-flow. (#809)

• App-friendly collected figures: theme / label / height hook (or pass a FigureSpec). (#801)

[2.1.1.post3] - 2026-07-30¶

Post-release of 2.1.1 — silence loader-discovery WARNING spam for apps.

• list_instruments() is quiet by contract: expected loader-probe skips (no DataLoader / missing custom def) log at DEBUG on the module logger; the scan uses create_all(quiet=True) so apps no longer see WARNING:root: Could not create loader … on every call. (#786)

[2.1.1.post2] - 2026-07-30¶

Post-release of 2.1.1 — collected summary facet y-axis controls for apps.

• Per-panel y-limits (and clearer share_y) for collected summary facet plots. (#804)

[2.1.1.post1] - 2026-07-29¶

Docs-only post-release of 2.1.1 — no API, schema, or runtime changes.

• Backfill HISTORY.md for the already-shipped 2.1.1 and 2.1.0.post1 releases. (#802)

[2.1.1] - 2026-07-29¶

Patch release — app-builder fixes and conveniences (gathered while building on 2.1.0.post1). Additive, no breaking changes.

• Bug fix (collect): group-averaged collected summaries plot again — Collection.plot() / plot(spread=True) render group-keyed mean±std frames instead of raising KeyError 'cell'. (#785)
• cellpy.collect.from_cells() (+ Batch.from_cells): build a batch from already-loaded in-memory CellpyCell objects. (#787)
• Collection.is_grouped / meta.grouped: whether group-averaging actually happened. (#790)
• Collection.save() supports xlsx and json (was parquet/csv only). (#789)
• CurveOptions gains mode / method, forwarded per cell to get_cap (parity with the single-cell call). (#788)
• cellpy.list_instruments(): quiet app-facing listing → [{"id", "label", "models", "suffixes"}], no per-module warnings. (#786)
• App/agents guide: note on the polars/pandas frame boundary and quieting cellpy's logger/warnings. (#791)

[2.1.0.post1] - 2026-07-28¶

Docs-only post-release of 2.1.0 — no API, schema, or runtime changes.

• Retire v2-docs-stable: docs land on master; RTD stable tracks release tags. (#775)
• Strip issue-tracker references and migration narrative from package docstrings so the API reference reads cleanly. (#771)
• Document Read the Docs header release-badge behaviour (badge shows the live GitHub latest release; the RTD version switcher is the source of truth). (#776)

[2.1.0] - 2026-07-28¶

cellpy 2.1 (Stage 4)¶

• Batch / collect redesign. New top-level cellpy.batch (journal / policy / runner / store / aggregate / outputs / facade) and cellpy.collect (options / collection / summary / curves / ica) packages replace the utils/batch_tools "farm/barn" machinery. cellpy.utils.batch and cellpy.utils.collectors remain as permanent re-export shims. (#697–#708, #716)
• Bug fix (collectors): cross-cell cycle collection no longer reassigns the shared cycles list — a cell missing a requested cycle no longer drops that cycle for every cell after it. (#705)
• New Batch.tests / aggregate.combine_tests. (#711)

• Breaking — 2.0 deprecation shims removed (Epic E). Every shim that warned "removed in 2.1" is gone; see docs/getting_started/migration_v2.0_to_2.1.md:

  • CellpyCell.headers_normal / _summary / _step_table → c.schema.raw / .summary / .steps; make_new_cell() → CellpyCell.vacant(). (#715)
  • ICA 1.x shims (Converter, dqdv_cycle / dqdv_cycles / dqdv_np, the legacy dqdv kwargs, the duplicate dq column) → ica.dqdv. (#714)
  • Plotting interactive=, xlim= / ylim=, summary_plot_legacy, and the seaborn + bokeh backends removed (plotly + matplotlib remain). (#713)
  • cellpy/utils/batch_tools/ deleted; the DB-journal path is now native in cellpy.batch. (#716)
  • The prms.* global-mutation shim (prms.Paths / prms.Reader / …) → cellpy.config (config.paths.x = … or config.override(...)). The CellpyCell.mass / .nom_cap property facades are kept. (#717)

• Example data regenerated as v9. The downloadable example cells served via cellpy.utils.example_data (CELLPY, OLD_CELLPY, RATE) are now v9 .cellpy (zip-of-parquet) files instead of v8 HDF5 .h5; same data, ~45% smaller. example_data.cellpy_file_path() / old_cellpy_file_path() / rate_file() now resolve .cellpy assets. (#718)

• Tried shipping marimo notebooks in the docs (#724); withdrawn — Zensical embeds could not keep a real marimo table/plot experience without hanging Pyodide or replacing widgets with non-marimo UI.

• Clarify Neware docs: binary .nda/.ndax via neware_nda (bundled fastnda) vs exported neware_txt models.

• Iterative docs fixes: example notebook typos/hedges, GITT intro quote, and batch config note for 1.x vs cellpy.toml. (#695).

• Loader golden oracles for host-local Arbin SQL datetime (arbin_sql_h5) compare relative offsets / omit start_datetime, so CEST regenerators and UTC CI agree (#768).

• Shipped to PyPI and conda-forge; docs synced to v2-docs-stable (#768).

[2.0.0] - 2026-07-26¶

• Stable cellpy 2.0.0. Native headers, v9 zip-of-parquet cellpy files (v8 HDF5 still readable; write v8 via escape), cellpycore==0.2.4, and the Stage-3 assembly close-out (#574). Soak tags: v2.0.0rc1, v2.0.0rc2.
• Support: Python ≥3.13. The v1.x line is bugfix-only for 12 months from this release date (decision #438-6). See docs/getting_started/migration_v1_to_v2.md and DEPRECATIONS.md.

• Known limitation (#687): scp:// / sftp:// URIs that use an OpenSSH Host alias (short name from ~/.ssh/config) may fail DNS after the OtherPath → UPath switch. Workaround: put the real hostname in the URI and set CELLPY_KEY_FILENAME — documented in docs/getting_started/remote_paths.md. Fix tracked in #687.

• Speed up remote auto_use_file_list / OtherPath.rglob dump. (#690).

• Follow directory symlinks in remote OtherPath.rglob so batch file discovery works under symlink project dirs. (#688).

• Rate-limit noisy collector warnings from max_segments interpolation fallback and dqdv half-cycle failures. (#669).

• Fix BatchICACollector fig_pr_cycle KeyError on cycle_num (use ICA cycle column). (#679).

• Add agent usage chapter for library/app builders. (#682).

• Update docs about data structure. (#680).

• Update docs and batch-loader examples for native step/raw headers (cellpy 2.0). (#676).

• Pin cellpycore==0.2.4 (EFC summary columns); sync HeadersSummary and pipeline_smoke goldens (#675).

• Iterative docs cleanup: landing pages, packaging/conda/dev guides, Mermaid and tutorial rendering, and a clean Zensical build (#673).

• Report dry-run setup destinations without claiming that files were written (#666).

• Align 2.0 release guides for a v2.0.0rc1 soak and record the Stage 3.17 gate audit (#574).

• Lock the file-format compatibility matrix (v8/v9 read/write, convert, pre-v8 freeze) (#573).

• Complete cli_api extraction for remaining CLI commands (new, serve, setup, …) (#651).

• Fix summary CV-split plots: use exclude_step_types and full−non_cv instead of dead selector_type (#654).

• Wire Batch.plot to plotting and delete batch_plotters.py (#658).

• Add ica_plot / dva_plot families on the new pipeline (#648).

• Port raw_plot and cycle_info_plot to prepare→spec→render (#647).

• Port cycles_plot to prepare→spec→render (#646).

• Matplotlib backend; retire SeabornPlotBuilder; unify backend= (#639).

• Port summary prepare path and flip summary_plot to prepare→spec→render (#638).

• Generic plotly panel/formation layout backend (#637).

• Add FigureSpec dataclasses and a PlotFamily registry (#636).

• Tier-3 loader close-out: biologics_mpr / batmo_bdf pass check_loader (native-projected adapters); ext_nda_reader parked; local_instrument confirmed as warn-only escape hatch (#561).

• Sync conda env files with pyproject.toml: pin cellpycore ==0.2.3 from conda-forge (was PyPI 0.2.1), add pyyaml / paramiko / universal-pathlib, drop obsolete fabric (#628).

• Include plotting tests in the nightly Tier-3 matrix: drop the stale test_plotutils_summary_plot --ignore from ci-scheduled.yml and set MPLBACKEND=Agg on conda-pytest / pip-install (#594).

• Docs: expand the 1.x→2.x migration guide (support matrix, config/c.schema, ICA, plotting, alpha caveats, loader notes) and point at DEPRECATIONS.md (#572).

• Maccor txt / 2.0.0a5: zero capacities from a swallowed pandas-3 Series.update failure (#580, fixed #581). Cellpy files saved from Maccor raw on 2.0.0a5 may have zeros baked in — re-load from raw after upgrading. Only maccor_txt (split_capacity: True) was affected.

• ICA redesign (#566): ica.dqdv() returns the long frame cycle, direction, voltage, capacity, dqdv (dq duplicate deprecated); new ica.dvdq(); failed half-cycles warn + frame.attrs["failures"]; old entry points shimmed to 2.1. Direction is cell-centric (#591).

• CLI: cellpy convert defaults to v9; --to {v9,v8} added; help is rich-formatted (Typer cutover #569). click is no longer a direct dependency.

• Plotting on 2.0.0a5: raw_plot / cycle_info_plot KeyError: 'voltage', summary_plot(x="cycle_index"), and in-memory summary corruption from CV-split y= sets are fixed (#593 / #567 Phase 0). Re-load in-memory cells that hit the CV-split bug; disk files were unaffected.

• Dependency budget (#570): drop required python-box / ruamel.yaml / python-dotenv; move PyTables to cellpy[legacy-files] (conda still ships pytables). Missing extra raises OptionalDependencyError with the install hint.

• Loaders: unknown undeclared vendor columns warn once (#599); Maccor model-one Watt-hr maps to energy (cumulative_charge_energy), not power (#599). harmonize() raises if a schema cast empties a column; new LoaderDeclarations.duration_columns for string durations (loader authors).

• Re-base collectors' drawing half onto cellpy.plotting (#657).

[2.0.0a6] - 2026-07-22¶

• Breaking: CellpyCell.get_cap now returns native cellpycore curve columns (#540, native-headers flip Stage 1): voltage → potential and cycle → cycle_num (capacity / direction unchanged). All in-repo consumers — plotutils.cycles_plot, the batch collectors, ica, and the CSV/Excel exporters — are updated; user code that indexes the get_cap frame directly must rename. See docs/getting_started/migration_v1_to_v2.md.

• Removed the deprecated cellpy.utils.easyplot module (#544). It was deprecated since 1.1 with removal scheduled for 2.0; use cellpy.utils.plotutils and cellpy.utils.collectors instead. The module, its tests, its docs entry, and its DEPRECATIONS.md row are gone.

• Step-type string literals use the cellpycore vocabulary (#543, native- headers Phase-0 prerequisite): CellpyCell.list_of_step_types is now list(config.STEP_TYPES) (was a hand-maintained duplicate of the 13 step-type names), and the step-table type comparisons in utils/ocv_rlx.py / utils/helpers.py use StepType.CHARGE.value / .DISCHARGE.value instead of bare "charge" / "discharge" literals. Behavior-identical.

• Header column literals in utils/helpers.py and filters/summary.py use header-object attributes (#538, native-headers Phase-0 prerequisite): the base-name string-keyed hdr_summary["charge_capacity"]-style lookups become attribute access, and filter_summary's rate_columns default is resolved from HeadersSummary (was a hard-coded ("charge_c_rate", "discharge_c_rate") tuple). Postfix/specific columns (*_gravimetric, areal_*) keep string-key composition. Behavior-identical.

• Journal-page column literals use HeadersJournal attributes (#537, native- headers Phase-0 prerequisite): the string-keyed hdr_journal["mass"]-style lookups in batch_plotters.py and helpers.py become attribute access (hdr_journal.mass, …) so a journal-header rename touches the header class only. Behavior-identical.

• Remote paths via universal_pathlib (#375, #371): OtherPath is now a thin wrapper around UPath (fsspec/Paramiko) instead of Fabric. Supported schemes remain ssh:// / sftp:// / scp:// (scp aliased to sftp). Remote exists / is_file / is_dir are truthful (no longer stubbed as always true). Saving a cellpy file to a remote URI raises a clear error. Credentials still come from CELLPY_KEY_FILENAME / CELLPY_PASSWORD. See docs/getting_started/remote_paths.md.

• Units are per-cell now (#427): get_cellpy_units returns a fresh CellpyUnits per call (optionally seeded from its argument, which used to be silently ignored), so changing units on one CellpyCell — directly, via the constructor, or via cellpy.get(units=...) — no longer changes them for every other cell in the session.

• Campaign merge supports renumber_cycles=False (#529, unblocked by cellpycore 0.2.2): sources keep their original cycle numbers — the identifying key becomes (test_id, cycle) and cycle-keyed consumers see the union of matching cycles; data points are still offset to stay globally unique. Steps/summary group and window per test, and the merged object round-trips through v9.

• Dependency-injection tail of V2-09 (#520): CellpyCell(core=..., instrument_factory=...) — the core seam and the loader registry are now constructor-injectable (defaults unchanged); register_instrument_readers() keeps an injected factory instead of silently rebuilding. ADR: .issueflows/04-designs-and-guides/cellpycell-di-restructuring.md.

• Split/drop-cycle helpers extracted from cellreader.py into cellpy.readers.slicing (#519, V2-09 follow-up): split, split_many, drop_from/drop_to, from_cycle/to_cycle, drop_edges, with_cycles, mod_raw_split_cycle moved verbatim (instance-first functions, thin delegates keep the public API); new pin tests added where coverage was thin.

• Exporter family extracted from cellreader.py into cellpy.exporters.tabular (#518, V2-09 follow-up): to_csv / to_excel and the _export_* helpers moved verbatim (instance-first functions, thin delegates keep the public API); the near-dead _cap_mod_* helpers moved along with their removal deferred to #520. A stray print(externals) debug line in to_excel was dropped.

• Native schema opt-in (#511, V2-11): CellpyCell(native_schema=True) keeps frames in native cellpy-core column names and runs the polars engine directly — no legacy rename sandwich. Supported pipeline: from_raw / load → make_step_table → make_summary → save (v9). Legacy path stays the default; legacy-named consumers (get_cap, exporters, plotting, campaign merge) are not supported on a native-schema cell yet.

• cellpycore 0.2.2 sync: the legacy bridge now carries test_id on steps and summary for all objects (core #136), so the #507 re-stamp workaround in make_step_table is removed and the campaign merger remaps the right-hand summary's test_id onto its new ids (previously the merged, pre-recompute summary showed test_id=0 everywhere).

• Cellpy-file format v9 (#510, V2-13/14): default save() writes zip-of-parquet + meta.json (.cellpy); load() still reads v4–v8 HDF5 and sniffs v9. Full TestMetaCollection (+ units/limits) persists on v9; campaign-merged multi-test_id objects round-trip. Escape: .h5 / cellpy_file_format="hdf5". Cellpy-owned save_meta_archive / load_meta_archive (core archive stubs stay stubs). User guide: docs/getting_started/migration_v1_to_v2.md.

• make_summary gains exclude_step_types (#509, v2 theme V2-12): step-type prefixes (e.g. ["cv_"]) whose per-cycle capacity contribution is subtracted from cycle-end charge/discharge capacities before derived columns — the core-native replacement (cellpy-core #54) for the removed selector-based exclusion. The deprecated no-effect exclude_types/exclude_steps/ selector_type/selector kwargs still only warn. Also fixes a latent load bug: older cellpy files with a double-nested cycle_mode ([['anode']]) are now unwrapped at the file-load boundary.

• Slim CellpyCell (#509, v2 theme V2-09): remove dead/experimental code — the _dev_update* family (broken call signatures, never wired), the module-level _check*/__main__ dev-scratch harness, _export_cycles_old, _select_steps, the long-dead select_steps/populate_step_dict (raised DeprecatedFeature since 1.x), and the superseded _select_without legacy exclusion. The capacity/curve getters (get_cap, get_ccap, get_dcap, get_ocv) moved verbatim to the new cellpy.readers.capacity_curves module with thin delegate methods — public API unchanged, curve goldens byte-identical. cellreader.py: 5987 → 4778 lines (−20%). Follow-ups tracked separately: exporter extraction, split/drop extraction, dependency-injection restructuring.

• Top-level API (#509, v2 theme V2-10): cellpy.get confirmed as the sanctioned entry point (stale removal-TODO dropped); cellpy.merge_cells and cellpy.print_instruments now exported at package level.

• Loaders emit per-test metadata (#508, v2 Phase 2, themes V2-05/06/08): from_raw now routes loader-parsed metadata (tester test_ID, channel_index, creator, schedule_file_name) into meta_test_dependent (so it persists and surfaces in Data.tests; the orphan attributes remain set for backward compatibility), stamps the compact test_id grouping key (0) onto raw (tester ids stay as provenance — note: this overwrites Arbin's per-row tester Test_ID values in the raw column), and records load provenance (uuid — new per load until #510 persists it, source_kind, source_type, source_uri, raw_file_names, loaded_datetime) that the derived TestMeta record now carries. Config-driven loaders stamp instrument raw_units on the returned Data (shared internal_settings.merge_raw_units helper). Arbin Global_Table.Comments now maps to meta_common.comment; the full vendor-column mapping (incl. deliberately dropped columns) is documented in the arbin_res module docstring. Loader goldens regenerated accordingly.

• Campaign merge (#507, v2 Phase 1-2, themes V2-03/V2-07): CellpyCell.merge is rewritten (old signature was dead code) — merge(cells, mode="campaign") folds different tests into one multi-test object: distinct compact test_id per source stamped on raw (overwrites tester-assigned ids; provenance stays in meta_test_dependent.test_ID), per-test metadata records in Data.tests, globally renumbered cycles, offset data points, unshifted timelines, and no cumulative carry-forward (summaries window per test on recompute). mode="continuation" keeps the classic fold (identical to from_raw([f1, f2]), which is untouched). New non-mutating merge_cells() helper and test_meta.cycle_ranges_per_test(). Step tables of campaign objects carry a test_id column so the engine groups and windows per test end-to-end. Also fixes latent bugs in _append's merge_step_table branch.

• Per-test metadata API (#506, v2 Phase 1, themes V2-01/02/04): Data.tests exposes a cellpycore.metadata.TestMetaCollection keyed by test_id (active record derived from the legacy meta boxes, which stay authoritative; extra records stored in memory), with Data.set_test_meta, Data.get_cycle_mode(test_id) / set_cycle_mode(mode, test_id) and Data.active_test_id. v1-v8 cellpy files load as a single-test collection (test_id=0). Engine compute on objects mixing different cycle_modes now raises MixedCycleModesError instead of silently applying one convention (per-test engine polarity: #507/#510). On-disk format v8 unchanged; only the active test's metadata is persisted (full collection persistence: #510). Adds the legacy<->core metadata mapping contract tests that cellpycore.legacy.meta_mapping assigns to cellpy. Future vocabulary/export alignment target: BattINFO (BIG-MAP).

• Header single source (#505, v2 Phase 0 gate): drop the redundant module-level HEADERS_NORMAL / HEADERS_SUMMARY / HEADERS_STEP_TABLE constants in cellreader.py and data_structures.py; use the instance attributes / get_headers_*() accessors from internal_settings instead.

• Decommission the in-repo legacy summary engine (#385, v2 Phase 0 gate): remove _make_summar_legacy, _generate_absolute_summary_columns, _ir_to_summary, _end_voltage_to_summary and the make_summary(old=True) branch — the cellpy-core engine is the only summary path. Breaking: the old kwarg is gone (1.x users keep it on the v1.x branch). The arbin_sql_h5 test now runs on the core path (verified value-identical; core prunes 13 duplicate raw rows, 47→34).

• Stage 3.3: single-file raw loads default to harmonize(parse()) (Reader.use_harmonized_raw=True); Arbin wide-aux columns keep values under aux_<quantity>_<name>; vendor datapoint_num is preserved; batmo_bdf decode runs in parse(); arbin_sql_h5 keeps all loader-stage rows. (#560)

1.1.0.post1 - 2026-07-15¶

• Sync conda env files (environment.yml, environment_dev.yml, github_actions_environment.yml) with pyproject.toml: pandas ≥3.0.3, cellpycore==0.2.1 from PyPI, drop obsolete pip tooling
• Remove invoke tasks.py; noxfile.py installs from pyproject.toml (.[all] + dependency-group dev)
• Docs: require Python 3.13+, copyright 2026; remove scratch root notes / Jupyter ZMQ fix markdown; refresh developer folder-structure tree

1.1.0 - 2026-07-15¶

• Pin cellpycore==0.2.1 (bridge honors cycle_mode for coulombic columns); regenerate pipeline_smoke goldens and fix native summary parity helper
• Stage 1.8: config migration — prms shim / legacy YAML path, migrate internal call sites, remove import-time config init (#453); cellpy setup writes cellpy.toml twin + setup migrate; info --config (#454)
• Stage 1.15 / 1.14: dormant native↔legacy frame translation (#458); Polars Phase A — de-index raw/summary/journal (#457)
• Stage 1.6: delegate duplicated unit converters to cellpycore.units (#451)
• Stage 1.7: parallel cellpy/config/ pydantic-settings stack — typed models, layered TOML loader, provenance, override(), inventory parity vs #430; not wired into legacy prms yet (#452)
• Fix: full-suite test failures — extract_fids test uses module helper; external check_file_ids test avoids live SCP; Arbin .res loader closes ODBC connections (#491)
• Stage 1.10: replace hard-coded column-header literals with canonical headers_* lookups in journal pages, ocv_rlx/plotutils, and instrument loaders (priorities 1–3); delete dead easyplot block (#455)
• Stage 1.4: redirect out-of-band HDF5 readers to cellpy_file.read_table / read_fid_table; CorruptCellpyFile for missing keys; cellpy convert CLI for v<8 upgrades (#449)
• Stage 1.3: move cellpy-file read/write paths into cellpy_file/ (#448)
• Units Phase 1: re-export CellpyUnits and Q from cellpycore; remove cellpy-local pint registry; rename cellreader data_structures alias to ds (#450)
• Deprecation: cellpy.utils.easyplot warns on import via warn_once; use plotutils/collectors instead (removed in 2.0, #438 decision 5) (#479)
• Fix: loader PEC golden compares all datetime columns by epoch-ns (Windows datetime64[us] vs ns); benchmark baseline gate warns above +20% slowdown and fails only above +100% (#476)
• Stage 1.1: extract cellpy-file format spec into cellpy/readers/cellpy_file/format.py; prms._cellpyfile_* aliases preserved; template registry and example-data URL constants moved to owning modules (#446)
• Stage 1.2: stateless cellpy-file helpers in cellpy_file/; explicit LoadSelector/LoadLimits replaces self.limit_* side channel during HDF5 extraction (#447)
• Stage 0 foundations complete — all linked characterization, oracle, baseline, convention, and decision-register issues closed; ready for Stage 1 (#439)
• Docs: Stage 0.11 decision register recorded in architecture-plan (timezone, curve-schema, v9 container, IR semantics, easyplot, v1.x maintenance) (#438)
• Conventions: cellpy._deprecation.warn_once helper, DEPRECATIONS.md registry, exception-tree stubs (CellpyError re-export plus CorruptCellpyFile, ConfigurationError, UnitsError, LoaderError), and make_new_cell wired as first consumer (#437, closes #456)
• Testing: legacy v4/v5 cellpy-file loads now covered in the v4–v7 characterization matrix (removed stale TypeError pin) (#466)
• CI: retry setup-miniconda in scheduled workflow on transient conda-forge download failures (#465)
• Testing: Stage 0.9 benchmark harness — opt-in pytest-benchmark suite under benchmarks/, committed v1.x baseline JSON, and dedicated CI job with ±20% regression gate (#436)
• Docs: Stage-0 AST inventory scanners (scan_member_usage.py, scan_hardcoded_headers.py) in .issueflows/00-tools/ for consumer/header reports (#435) (tests/parity.py::assert_value_parity) — legacy vs native frames compared through cellpycore.legacy.mapping with dtype-tolerant mapped-column equality, named exception list, and trivial-pass essential tests on the current bridge (#434)
• Testing: curve-extraction golden snapshots for get_cap / get_ccap / get_dcap / get_ocv on the canonical Arbin cell, including labeled/interpolated/multi-cycle variants and NullData edge cases (#433)
• Testing: per-loader golden snapshots for tier-1 loaders (arbin_res, maccor_txt, neware_txt, pec_csv, custom) — raw frame, raw_units, and loader meta oracles under tests/data/goldens/loader_*/ with parametrized essential regression tests (#432)
• Testing: prms configuration characterization tests — inventory parity contract, config round-trip/precedence, OtherPath coercion, .env_cellpy pickup, and cellpy setup dir/file creation (#430)
• Testing: cellpy-file HDF5 characterization tests — v8 round-trip (fid-populated fixture), limits-prefix trap, max_cycle selector, legacy version matrix, and missing-key failure mode (#429)
• Testing: golden-fixture convention under tests/data/goldens/, dev/regenerate_goldens.py suite registry, and pipeline_smoke essential oracle on the canonical Arbin .res (#428)
• Docs: clarify cellpy_units defaults in cellpy.get() and Google docstring formatting cleanup in cellreader.py (#425)
• Fix: bump pandas to 3.0.3 — BDF Unix-time export, batch journal string extraction, and post-processor datetime handling (#415)
• CI: move Windows conda pytest from AppVeyor to GitHub Actions (ACE x64 install with cache) (#407)
• Testing: new offline unit tests for utils/helpers.py (outlier removal, group names, rate column) and readers/filefinder.py (tmp-path raw-file trees) (#372)
• Fix: local OtherPath.rglob() did not recurse into subdirectories, so search_for_files(..., sub_folders=True) missed files in subfolders (#372)
• Fix: remove_outliers_from_summary_on_nn_distance crashed on pandas ≥ 2 (TypeError in 2-element window branch) (#372)
• Fix: list_raw_file_directory(extension=...) raised TypeError when filtering path objects (#372)
• Testing: silenced easyplot SyntaxWarning and corrected an xfail marker to use raises= (#372)

1.0.4a1 - 2026-07-03¶

First alpha on the automated GitHub release → PyPI pipeline. Install with pip install cellpy --pre or pip install cellpy==1.0.4a1.

• Integration: cellpycore consumed from PyPI (pinned 0.1.2 for this release); core step/summary processing delegated via OldCellpyCellCore seam (#377, #400–#401)
• CI: release.yml — published GitHub release triggers test + PyPI trusted publishing (#403)
• Testing: essential pytest marker for read → step-table → summary smoke + parity contract
• Build: Python ≥ 3.13; hatchling + uv-dynamic-versioning from git tags (#354)
• Removed dependency on cellpy-core's deprecated create_selector (#399)
• Plus fixes and features merged since v1.0.3a6 (PEC reader #393, batch/config #392/#397, module rename #381, filters #363, BDF export #356, …)

1.0.3 (pre-release)¶

• Refactor: Renamed internal modules accidentally named core — cellpy.readers.core → cellpy.readers.data_structures and cellpy.internals.core → cellpy.internals.connections — to avoid a name clash with the new cellpy-core package (#381)
• Testing: Added header/unit parity contract tests asserting cellpy-core's settings copies (HeadersNormal, HeadersSummary, HeadersStepTable, CellpyUnits) stay in sync with cellpy's internal_settings (#378)
• Build: Migrated packaging to a single pyproject.toml (hatchling + git-tag dynamic versioning), managed with uv; removed setup.py/requirements*.txt/MANIFEST.in and added a Docker-based local build test (#354)
• Filters: add filtering possibility to plotters in plotutils (#363)
• Fix: clear leaky seaborn facet titles (row = ... | cycle_type = standard) in multi-row summary_plot panes when show_formation=True, while preserving x-tick labels on the bottom row (pre-existing bug surfaced by the new _with_rate y-sets)
• Exporters: New CellpyCell.to_bdf(...) exports raw time-series in Battery Data Format (CSV/Parquet) with optional cycle filtering, for use by UiA's dsToolbox and other BDF-aware tools (#356)
• Exporters: to_bdf now accepts an extras keyword for appending custom/auxiliary raw columns alongside the BDF payload (extras=True for all unmapped columns, or pass a list/string of column names). Extras are written verbatim with no unit conversion; the resulting file is no longer strictly BDF-compliant.
• General: New cellpy.exporters and cellpy.filters packages; the CellpyCell class layer no longer imports from cellpy.utils
• Batch: Batch plotting with multiple subfigures (#343, #344, #346)
• Batch: JSON db reader from batbase (batbase_json_reader) - new database reader for JSON-based batch files
• Batch: Improved batch load functionality
• Batch: Enhanced error handling and logging in batch processing with clearer exception messages
• Batch: Changed default output directory name from 'out' to 'dump'
• Batch: Use local folder for journal file as default
• Batch: Allow prms to pass during batch update when reloading cellpy files
• Batch: Improved concat_summaries with support for different averaging methods and filtering (low/high values)
• Batch: Added CV-share partitioning support in summary collector
• Batch: Added line hooks in summary plot
• Batch: Summary plot now supports fullcell standard
• Batch: Added possibility to drop columns and filter low/high for non-grouped data
• Batch: Added helper function collectors.standard_gravimetric_collector
• General: Require numpy >= 2
• General: Made explicit imports of parameters and readers in top init-file
• General: Allow additional arguments to plotly save images
• CLI: New label for create new projectdir in cellpy new
• Readers: JSON db reader now supports optional storage of raw JSON data via store_raw_data parameter
• Readers: Added raw_pages_dict and pages_dict properties to JSON db reader for accessing data as dictionaries
• Bug fixes: Fixed bug in pandas.ExcelWriter call (#347)
• Bug fixes: Fixed bug in summary collector (concat summaries) that mutated list of selected columns
• Bug fixes: Fixed bug in OtherPathsNew
• Bug fixes: Various other bug fixes and improvements
• CI: Fix failing CI pipelines (pyarrow runtime dep + AppVeyor 64-bit Miniconda) (#360)
• Bug fixes: Fix TypeError: bad operand type for unary ~: 'slice' in plotutils.summary_plot when called with formation_cycles=False or 0 (#366)
• Exporters: to_bdf accepts a bdf_units keyword to control units written into the BDF file (#365)

1.0.2¶

• Batch: only_selected keyword added for concatenating summaries choosing only selected cells in the pages (selected==1)
• General: Add option to specify custom_log_path and path to logging config json in get() (#326) by @morrowrasmus
• Batch: implement wide format for collectors to csv
• Batch: adding more columns to pages (model, selected, nom_cap_specifics)
• General: Implemented lazy import to speed up loading of cellpy
• General: Added _absolute cols in the summary
• General: Add basic support for reading parquet for custom instruments (#322) by @morrowrasmus
• Utils: General improvements in plotutils
• General: Dropped support for python 3.9 and added support for python 3.12 (and probably beyond) by upgrading OtherPaths
• Bug fixes.

1.0.1¶

• Utils: example_data now includes auto-download of example data
• General: supports only python 3.10 and up to 3.11
• Batch: naked and init(empty=True) easier method for creating batch with empty pages
• File handling: new fix in find_files
• Batch / Utils: refactored and updated Collectors (using plotly)
• Batch: new summary plotter (using plotly)
• Batch: new convenience function for automatically creating batch from batch-file if file exists.
• Batch: added mark and drop methods
• CLI: added possibility to use custom jupyter executable
• Added checks (c.has_xxx) for checking if data has been processed correctly / fix errors in raw/semi-processed data.
• Added possibility to filter on C-rates (c.get_cycles)
• Added experimental feature c.total_time_at_voltage_level for calculating total time at low/high voltage
• Added experimental instrument reader for neware xlsx files (hopefully not used much because it is very slow)
• Added try-except block for ica post-processing step and add if-clause (suggested by Vajee)
• Fixed several smaller bugs and improved some of the functionality (most notably in c.get_cap)
• Added CI for macOS
• Added conda package including sqlalchemy-access
• Improved plotting tools
• Improved documentation
• Improved feedback from the CLI

1.0.0 (2023)¶

• Unit handling: new unit handling (using pint)
• Unit handling: renaming summary headers
• Unit handling: new cellpy-file-format version
• Unit handling: tool for converting old to new format
• Unit handling: parsing input parameters for units
• Templates: using one repository with sub-folders
• Templates: adding more documentation
• File handling: allow for external raw files (ssh)
• Readers: neware.txt (one version/model)
• Readers: arbin_sql7 (experimental, @jtgibson91)
• Batch plotting: collectors for both data collection, plotting and saving
• OCV-rlx: improvements of the OCV-rlx tools
• Internals: rename main classes (CellpyData -> CellpyCell, Cell -> Data)
• Internals: rename .cell property to .data
• Internals: allow for only one Data object pr CellpyCell object
• CLI: general improvements and bug fixes
• CLI: move editing of db-file to the edit sub-command

0.4.3 (2023)¶

• Neware txt loader (supports one specific format only, other formats will have to wait for v.1.0)

0.4.2 (2022)¶

• Changed definition of Coulombic Difference (negative of previous)
• Updated loaders with hooks and additional base class TxtLoader with configuration mechanism
• Support for Maccor txt files
• Supports only python 3.8 and up
• Optional parameters through batch and pages
• Several bug fixes and minor improvements / adjustments
• Restrict use of instrument label to only one option
• Fix bug in example file (@kevinsmia1939)

0.4.1 (2021)¶

• Updated documentations
• CLI improvements
• New argument for get_cap: max_cycle
• Reverting from using Documents to user home for location of prm file in windows.
• Easyplot by Amund
• Arbin sql reader by Muhammad

0.4.0 (2020)¶

• Reading arbin .res files with auxiliary data should now work.
• Many bugs have been removed - many new introduced.
• Now on conda-forge (can be installed using conda).

0.4.0 a2 (2020)¶

• Reading PEC files now updated and should work

0.4.0 a1 (2020)¶

• New column names (lowercase and underscore)
• New batch concatenating and plotting routines

0.3.3 (2020)¶

• Switching from git-flow to github-flow
• New cli options for running batches
• cli option for creating template notebooks
• Using ruamel.yaml instead of pyyaml
• Using python-box > 4
• Several bug-fixes

0.3.2 (2019)¶

• Starting fixing documentation
• TODO: create conda package
• TODO: extensive tests

0.3.1 (2019)¶

• Refactoring - renaming from dfsummary to summary
• Refactoring - renaming from step_table to steps
• Refactoring - renaming from dfdata to raw
• Refactoring - renaming cellpy.data to cellpy.get
• Updated save and load cellpy files allowing for new naming
• Implemented cellpy new and cellpy serve cli functionality

0.3.0 (2019)¶

• New batch-feature
• Improved make-steps and make-summary functionality
• Improved cmd-line interface for setup
• More helper functions and tools
• Experimental support for other instruments
• invoke tasks for developers

0.2.1 (2018)¶

• Allow for using mdbtools also on win
• Slightly faster find_files using cache and fnmatch
• Bug fix: error in sorting files when using pathlib fixed

0.2.0 (2018-10-17)¶

• Improved creation of step tables (much faster)
• Default compression on cellpy (hdf5) files
• Bug fixes

0.1.22 (2018-07-17)¶

• Parameters can be set by dot-notation (python-box).
• The parameter Instruments.cell_configuration is removed.
• Options for getting voltage curves in different formats.
• Fixed python 3.6 issues with Read the Docs.
• Can now also be used on posix (the user must install mdb_tools first).
• Improved logging allowing for custom log-directory.

0.1.21 (2018-06-09)¶

• No legacy python.

0.1.0 (2016-09-26)¶

• First release on PyPI.