Feature/phonon reporting - #20
Open
alinzh wants to merge 14 commits into
Open
Conversation
alinzh
commented
Sep 1, 2026
alinzh
commented
Collaborator
- Add phonon processing
- Add actual Vultr pricing
…e overrides - Add bare-metal plans (vbm-*) to CLOUD_PRICING from Vultr API - Add detect_provider() vbm token, read_provider_and_machine_type_from_config() - Add --provider, --engine, --machine-type CLI flags to dft-report-aiida/dft-report - Rename hetzner_rate->cloud_rate, cost_eur->cost, add currency field - provider+machine_type fallback: CLI flag -> /etc/yascheduler/yascheduler.conf -> skip - When no provider info available (no flag, no config), cost columns are omitted from CSV - --engine filter skips expensive FLEUR enrichment (Seebeck/displacement) - Remove dead code in pricing.py (duplicate lines after return) - 62 tests: bare-metal rates, config read, skip_cost, provider/engine/machine_type overrides Closes #17
All tasks complete (62 tests pass, verified on real AiiDA data). Moved to openspec/changes/archive/2026-08-04-add-vultr-cost-calculation/
- Fix amd5 substring collision: longest-key-first matching in get_cloud_rate - Unknown provider/plan returns None instead of 0.0 (no silent 'free' cost) - UpCloud recognised in config but has no pricing → rate/currency None - resolve_provider_and_rate(): single shared helper, validates provider+machine_type consistency (no cross-cloud mixing, e.g. vultr+ccx13 → rate None) - Remove skip_cost: auto-detect from computer name still works without config - Currency set even when rate is None (provider known, plan unknown) - Config tests use tmp_path (no /etc/yascheduler/ dependency) - Engine filter tests verify QueryBuilder.append() gets process_type filter - 81 tests: amd5, cross-provider mismatch, unknown plans, no-config auto-detect Addresses PR #19 review feedback.
… integration Merge CRYSTAL phonon parsing from PR #18 with FLEUR phonon WorkChain enrichment and thermodynamic integration (ZPE, F, S, Cv) from ab_initio_calculations.phonon_thermo. CRYSTAL (from PR #18): - crystal_parser/properties/phonon.py: parse phonon frequencies from OUTPUT MODES block, PHONON.DAT, FREQ.DAT - _enrich_crystal_extras: has_phonons, phonon_freq_min/max/mean/std, phonon_n_imag, phonon_modes_count - determine_calc_type_summary: phonon/transport/elastic/electron types - --calc-type, --max-duration, --skip-displacement CLI flags FLEUR (new): - phonon_utils.py: extract frequencies from PhonopyFleurWorkChain via phonopy force_constants, integrate via custom/phonopy/ase methods - _enrich_phonon_data: enrich phonon CalcJobs by walking provenance to parent PhonopyFleurWorkChain - n_imaginary, zpe_kjmol, f_at_t_kjmol, s_at_t_jkmol, cv_at_t_jkmol, phonon_n_qpoints, phonon_n_bands, t_eval columns - dft-report-phonons CLI: standalone phonon table with PR, SG, Pearson, formula, cell params, cost, thermodynamic properties - --phonon-t-eval, --phonon-method CLI flags Shared: - structures.py: get_space_group_robust, nullify_right_angles - _pearson_from_attrs: Pearson symbol filling in _enrich_with_structure_fast - resolve_provider_and_rate (Vultr+Hetzner) replaces get_cloud_rate/get_cost in AiiDA-DB pipeline - ab_initio_calculations dependency for phonon_thermo integration
dft-report-phonons now queries both FLEUR PhonopyFleurWorkChain nodes and CRYSTAL phonon CalcJobNodes, producing a unified phonon table. CRYSTAL phonon frequencies are parsed from the retrieved OUTPUT file (MODES block), grouped into q-points, and integrated via the same thermodynamic methods (custom/phonopy/ase) as FLEUR. New functions: - extract_frequencies_from_crystal_calc(): parse OUTPUT → freqs cm⁻¹ - get_crystal_phonon_summary(): full summary for CRYSTAL phonon CalcJob - scan_phonon_workchains() now queries both WorkChainNode + CalcJobNode Added 'engine' column to distinguish fleur/crystal in the phonon CSV. Added CRYSTAL phonon columns (has_phonons, phonon_freq_min/max/mean/std, phonon_n_imag, phonon_modes_count) to _PHONON_CSV_COLUMNS. Fixed: label filter for CRYSTAL phonon query (AiiDA QueryBuilder does not support 'like' on label column — post-query filtering instead).
Main report (dft-report-aiida) now adds only two phonon columns: - phonon_pk: PK of the phonon calculation (WorkChain for FLEUR, CalcJob for CRYSTAL) - has_phonon: boolean flag Full thermodynamic properties (ZPE, F/S/Cv, n_imaginary, etc.) are available via the standalone dft-report-phonons CLI. Removed phonon_t_eval, phonon_method, phonon_mesh parameters from scan_aiida_calculations, generate_aiida_reports, and report_aiida_cli.
Priority for space_group filling: 1. Label (e.g. 'ZnO/186: Phonon frequencies' -> 186) — conventional SG 2. mpds_query.sgs — conventional SG from MPDS 3. spglib on StructureData — only if label/mpds have no SG Previously spglib ran first on the primitive cell (phonopy supercell) and gave wrong conventional SG (e.g. P1 instead of P-3m1 for ZnO/186). Now label is parsed in initial summary, before any spglib enrichment. Also fixed _sgs_from_workchain_label to strip trailing ':' from tokens (e.g. '186:' in 'ZnO/186: Phonon frequencies [1]'). Removed pearson from main report (_SUMMARY_CSV_COLUMNS) — it was computed on the primitive cell and gave wrong values. Pearson remains available in dft-report-phonons standalone CLI.
FLEUR WorkChain labels (e.g. 'MgO/225 - phonons') and CRYSTAL labels (e.g. 'ZnO/186: Phonon frequencies [1]') contain the conventional space group number. Now parsed first, with spglib on the structure as fallback. Fixes: MgO showed SG=160 (rhombohedral primitive) instead of 225 (Fm-3m), KNbO3 showed 160 instead of 221 (Pm-3m), ZnO showed 1 instead of 186.
Full report: removed has_phonons, phonon_freq_*, phonon_n_imag, phonon_modes_count, cloud_rate — keeping only phonon_pk + has_phonon for phonon rows. Phonon report: removed has_phonons, phonon_freq_*, phonon_n_imag, phonon_modes_count, cloud_rate — these were CRYSTAL-specific and redundant with n_imaginary/n_qpoints/n_bands. Fixed FLEUR cost/currency/duration: WorkChainNode has no computer, so _find_remote_computer() walks children to find the yascheduler computer. All 19 phonon rows now have cost/currency/duration filled.
…lumns FLEUR PhonopyFleurWorkChain nodes are now added as separate rows in the main report with calc_type='phonon', engine='fleur'. This replaces the phonon_pk/has_phonon columns — phonon calculations are simply rows like scf/optimise/transport. Removed phonon_pk and has_phonon from _SUMMARY_CSV_COLUMNS and _null_summary_keys. _enrich_phonon_data -> _enrich_phonon_computer: only fills computer/cost/currency for FLEUR phonon rows (WorkChainNode has no computer, so _find_remote_computer walks children to find yascheduler). Fixed _formula_from_label to handle labels without ':' (e.g. 'ZnO/186 - phonons' -> 'ZnO' via '/' fallback).
Reordered _SUMMARY_CSV_COLUMNS and _PHONON_CSV_COLUMNS for readability: identity fields first (formula, pk, uuid, calc_type, engine, SG), then metadata (label, date, computer, exit), then results (duration, cost, bandgap, energy, etc.), then cell params last.
scan_calculations now returns (summary_store, error_dict_crystal, error_dict_fleur, phonon_store) — 4 values. archive_and_save expected only 3, causing ValueError in aiida-mpds-monitor archive generation. Fix: use *_ to absorb the extra phonon_store.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.