When a genus+species abbreviation like "P. ovale" was used, the previous
bypass (Issue #288) checked sum(sp_exact) == 1, which failed if the species
also had subspecies sharing the epithet (ovale curtisi, ovale wallikeri).
The fix extends the bypass to fire whenever all exact species matches belong
to one genus, collapsing to the species-rank record (subspecies == "") for
genus+species queries and preserving the chosen row for explicit subspecies
queries.
Also extends the data-invariant test to cover all taxonomic rank columns
from domain to subspecies, not just the terminal three.
Claude-Session: https://claude.ai/code/session_01M4fqQYQYJ3drdudkDYNqAY
Co-authored-by: Claude <noreply@anthropic.com>
* docs: rename remaining 'antibiotic selectors'/'AB selectors' to 'antimicrobial selectors'/'AMR selectors'
Five leftover occurrences of the old terminology updated in the AMR
vignette (section heading + code comment), the amr_selectors test file,
and two roxygen example comments in amr_selectors.R. The auto-generated
man/antimicrobial_selectors.Rd will update on the next devtools::document() run.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0169LucQ7SHDLHdTnDs1ENhd
* fix
---------
Co-authored-by: Claude <noreply@anthropic.com>
* Generalise interpretive rules for multi-guideline support (#268)
- Rename data-raw/eucast_rules.tsv → interpretive_rules.tsv; add rule.provider
column (value: "EUCAST") to distinguish future CLSI rows
- Rename EUCAST_RULES_DF → INTERPRETIVE_RULES_DF in _pre_commit_checks.R;
filter by rule.provider == guideline when applying rules in interpretive_rules()
- Rename custom_eucast_rules() → custom_interpretive_rules() with new S3 class
"custom_interpretive_rules"; old function becomes a deprecated wrapper in
zz_deprecated.R; backward-compat S3 dispatch shims added for old class
- Remove stop_if(guideline == "CLSI", ...) so clsi_rules() no longer errors
- Add .onLoad shim in zzz.R to create INTERPRETIVE_RULES_DF from EUCAST_RULES_DF
for transitional compatibility until sysdata.rda is regenerated
https://claude.ai/code/session_01D46BTsfJSPo3HnLWp3PRkP
* Fix namespace load failure: remove assignInNamespace from .onLoad (#268)
assignInNamespace cannot add NEW bindings to a locked package namespace
(R locks namespace bindings before .onLoad runs). Replace the .onLoad
shim with a runtime fallback inside interpretive_rules(): if
INTERPRETIVE_RULES_DF is absent (pre-regeneration sysdata.rda), derive
it from EUCAST_RULES_DF by adding the rule.provider column. This also
fixes the screening_abx line to reuse the already-resolved
interpretive_rules_df_total instead of a bare INTERPRETIVE_RULES_DF
reference.
https://claude.ai/code/session_01D46BTsfJSPo3HnLWp3PRkP
* fixes
* fixes
---------
Co-authored-by: Claude <noreply@anthropic.com>
* Add parallel computing support to antibiogram() and wisca() (#281)
For WISCA: simulations are distributed across (group, chunk) job pairs
via future.apply::future_lapply(), keeping all workers active even when
the regimen count is smaller than nbrOfWorkers(). Sequential fallback
with progress ticker is preserved when parallel = FALSE or workers = 1.
For grouped antibiograms: each group is processed by a separate worker,
mirroring the row-batch approach in as.sir().
Same gate pattern as as.sir() (PR #280): requires a non-sequential
future::plan() to be active; auto-upgrades to parallel = TRUE when a
parallel plan is detected; throws an informative error otherwise.
https://claude.ai/code/session_01FC43syPbzhGmKgrrVNHjnF
* Fix version to 3.0.1.9055 and update CLAUDE.md version formula
Uses origin/${defaultbranch} (with a fetch) instead of the local
branch ref so the commit count is never stale after a merge.
https://claude.ai/code/session_01FC43syPbzhGmKgrrVNHjnF
* Fix non-ASCII characters in antibiogram.R
Replace en/em dashes and non-breaking spaces with ASCII equivalents
to satisfy R CMD check portability requirement.
https://claude.ai/code/session_01FC43syPbzhGmKgrrVNHjnF
* Update auto-generated Rd files after documentation rebuild
https://claude.ai/code/session_01FC43syPbzhGmKgrrVNHjnF
* Move parallel gate to top of antibiogram.default() like sir.R
The gate was inside the wisca==TRUE block, so parallel=TRUE with a
sequential plan was silently ignored for non-WISCA antibiograms.
Now the gate runs unconditionally at the top of the function,
identical to the as.sir() pattern: error on explicit parallel=TRUE
with sequential plan, auto-upgrade when a non-sequential plan is
already active.
https://claude.ai/code/session_01FC43syPbzhGmKgrrVNHjnF
* Fix parallel WISCA returning all NA; strengthen tests; add sequential hint
Bug: lapply() over a factor yields length-1 factor elements (integer
codes), while for() over a factor yields character strings. The job
list stored j\$group as a factor integer, but the reassembly loop
compared it with identical(j\$group, g) where g was character -- always
FALSE, so no simulation chunks were ever assembled and coverage stayed
NA throughout.
Fix: convert unique_groups to character before building jobs so both
the job list and the reassembly loop use the same type.
Tests: replaced na.rm = TRUE guards with explicit anyNA() checks so the
test suite would have caught the all-NA result immediately.
Also adds a sequential-mode performance hint (analogous to sir.R
lines 1116-1127) when simulations >= 500 and >= 3 regimens.
https://claude.ai/code/session_01FC43syPbzhGmKgrrVNHjnF
---------
Co-authored-by: Claude <noreply@anthropic.com>
* Migrate parallel computing in as.sir() from parallel:: to future/future.apply
Replace parallel::mclapply() and parallel::parLapply() with
future.apply::future_lapply(), enabling transparent support for any
future backend (multisession, multicore, mirai_multisession, cluster)
on all platforms including Windows.
When parallel = TRUE the function now: (1) respects an active
future::plan() set by the user without overriding it on exit, or
(2) sets a temporary multisession plan with parallelly::availableCores()
and tears it down on exit. The max_cores argument controls worker count
only when no user plan is active.
future and future.apply are added to Suggests in DESCRIPTION.
https://claude.ai/code/session_01M1Jvf2Miu6JL4TQrEh1wS8
* Require user plan() for parallel=TRUE; fix as_wt_nwt false-positive warnings
- parallel = TRUE now errors with a cli-styled message if no non-sequential
future::plan() is active; users must call e.g. future::plan(future::multisession)
before using parallel = TRUE (breaking change)
- Removed auto-setup/teardown of multisession plan inside as.sir(), which was
slow and caused version-mismatch issues with load_all() workflows
- Added as_wt_nwt to the exclusion list in as_sir_method() to suppress
false-positive "no longer used" warnings during parallel runs
- Fixed pieces_per_col row-batch calculation to use n_workers (total available
workers from the active plan) instead of n_cores (workers clipped to n_cols),
so row-batch mode activates correctly when n_cols < n_workers
- Updated @param parallel and @param max_cores roxygen docs; regenerated man/as.sir.Rd
- Updated sequential-mode hint to instruct users to set plan() first
https://claude.ai/code/session_01M1Jvf2Miu6JL4TQrEh1wS8
* fix parallel
* fix parallel
* unit tests
* unit tedts
---------
Co-authored-by: Claude <noreply@anthropic.com>
* Fix custom reference_data support in as.sir() (#239)
- custom guideline names now correctly classify values as R: CLSI convention
(>= breakpoint_R for MIC, <= for disk) applies only when guideline contains
"CLSI"; all other guidelines including custom ones use the EUCAST convention
(> breakpoint_R for MIC, < for disk)
- guideline argument is now optional when reference_data is manually set: if
omitted or if its value does not match any row in the custom data, all rows
in reference_data are used; if set to a value present in the data, only
matching rows are filtered — useful for multi-guideline custom tables
- host = NA in custom reference_data now acts as a host-agnostic fallback
when no host-specific breakpoint row exists for the current animal species
- updated reference_data argument documentation to explain these conventions
https://claude.ai/code/session_01Q8KtFFGG9qrjAgLJBbxG2U
* Refactor R-classification logic using custom_breakpoints_set flag
Introduce custom_breakpoints_set <- !identical(reference_data, AMR::clinical_breakpoints)
at the top of as_sir_method() and replace all identical() calls inside that
function with this variable.
In the case_when_AMR interpretation blocks (MIC and disk), the R-classification
now has three explicit arms:
- !custom_breakpoints_set & EUCAST guideline -> open interval (> / <)
- !custom_breakpoints_set & CLSI guideline -> closed interval (>= / <=)
- custom_breakpoints_set -> open interval (> / <), always,
regardless of the guideline name in the custom data (e.g. "CLSI_custom"
must not accidentally trigger CLSI convention)
https://claude.ai/code/session_01Q8KtFFGG9qrjAgLJBbxG2U
* Fix unit tests for custom reference_data (#239)
- Do not override my_bp$mo / my_bp$ab in tests: assigning plain character
strips the <mo>/<ab> class, which check_reference_data() rejects. Use the
mo/ab values already present in the source row instead.
- Use NA_character_ instead of NA for my_bp$host so the host column keeps
its character class.
- Pass breakpoint_type = "animal" explicitly in the host-fallback test since
the custom reference_data only contains animal-type breakpoints.
https://claude.ai/code/session_01Q8KtFFGG9qrjAgLJBbxG2U
* Add coerce_reference_data_columns() for lenient reference_data validation
check_reference_data() now returns the (possibly coerced) reference_data and
the call site captures the result so downstream code sees the fixed columns.
A new coerce_reference_data_columns() helper is called before the strict class
check inside check_reference_data(). It coerces columns to the expected types:
- mo -> as.mo() if not already <mo> class
- ab -> as.ab() if not already <ab> class
- character columns -> as.character() (e.g. host = NA becomes NA_character_)
- numeric columns -> as.double()
- logical columns -> as.logical()
This allows users to build a custom reference_data from a plain data.frame
without having to pre-apply as.mo()/as.ab() or worry about NA column types.
Updated the reference_data roxygen argument to document the auto-coercion and
restored the tests to the simpler form that uses plain character assignments,
relying on the new coercion instead of workarounds.
https://claude.ai/code/session_01Q8KtFFGG9qrjAgLJBbxG2U
---------
Co-authored-by: Claude <noreply@anthropic.com>
* Fix parallel computing in as.sir.data.frame
Six bugs in parallel = TRUE mode:
1. PSOCK workers (Windows / R < 4.0) never had AMR loaded, so every
exported/AMR function call failed. Added clusterEvalQ(cl, library(AMR))
with a graceful fallback to sequential when the package cannot be loaded
(e.g. dev-only load_all() environments).
2. clusterExport'd AMR_env was a frozen serialised copy; as.sir() on the
worker wrote to AMR:::AMR_env while run_as_sir_column read from the stale
copy, so the captured log was always wrong. Fixed by resolving AMR_env
dynamically via get("AMR_env", envir = asNamespace("AMR")) inside the
worker function, and removing AMR_env from clusterExport.
3. In the fork-based (mclapply) path each worker inherited the parent's full
sir_interpretation_history. Capturing the whole log then combining across
workers duplicated every pre-existing entry. Fixed by recording the log
row count before the as.sir() call and slicing only the new rows
afterwards.
4. run_as_sir_column used non-exported internals (%pm>%, pm_pull,
as.sir.default) that are inaccessible on PSOCK workers after library(AMR).
Replaced pipe chains with direct as.mic(as.character(x[, col, drop=TRUE]))
and as.disk(...) calls, and changed as.sir.default() to as.sir() which
dispatches correctly via S3.
5. With info = TRUE, worker forks printed per-column progress messages
simultaneously, producing garbled interleaved console output. Per-column
messages are now suppressed inside workers (effective_info = FALSE) while
the outer "Running in parallel" / "DONE" messages still appear.
6. Malformed Unicode escape \u00a (3 hex digits) in the "DONE" banner was
parsed by R as U+00AD (soft hyphen) + "ONE"; corrected to .
https://claude.ai/code/session_012DXCXbZUC54Zij1z9bFiHR
* Add parallel computing tests to test-sir.R
Eight targeted tests verify correctness of the parallel as.sir() path:
identical SIR output vs sequential, matching log row counts, no
pre-existing history duplication, reproducibility across runs, results
consistency across max_cores values, single-column fallback, and no
per-column worker messages leaking when info = TRUE. All pass when only
1 core is available (parallel silently falls back to sequential).
https://claude.ai/code/session_012DXCXbZUC54Zij1z9bFiHR
* Fix as.sir() data.frame: preserve already-<sir> columns, exclude metadata
Issue #278: two related bugs in the column-detection / type-assignment pipeline.
Bug 1 – already-<sir> columns deleted on re-run
Line 886 excluded already-sir columns from the type assignment (they
stayed type "") causing the result loop to do x[,col] <- NULL, deleting
them. Fix: drop the !is.sir() guard so all untyped columns fall through
to type "sir" and are re-processed correctly.
Bug 2 – metadata columns treated as antibiotics
as.ab("patient") -> OXY, as.ab("ward") -> PRU. The column detector
accepted any column whose name matched an antibiotic code, regardless of
content. Fix: for name-matched columns that do not already carry an AMR
class, also verify content looks like AMR data (all_valid_mics, all-
numeric, or any SIR-like string). all_valid_disks() is intentionally
avoided here because it strips letters from strings (as.disk("Pt_1")==1).
Also adds tools/benchmark_parallel.R: a standalone script that times
sequential vs parallel as.sir() across n=20/200/2000/20000 rows and
saves a ggplot2 PNG to tools/benchmark_parallel.png.
https://claude.ai/code/session_012DXCXbZUC54Zij1z9bFiHR
* Update benchmark: two-panel script with warm-up and column-count sweep
Previous single-panel benchmark was misleading: the first sequential run
paid one-time cache-warm-up cost (skewing n=20), and only 6 columns were
used so only 6 cores were ever active on a 16-core machine.
New two-panel design:
Left – vary rows with 16 fixed AB columns (shows memory-bandwidth
saturation for large n)
Right – vary columns with fixed rows (shows the real speedup profile:
parallel wins when n_cols >> 1)
Also adds a warm-up pass before measurements to eliminate first-call bias.
https://claude.ai/code/session_012DXCXbZUC54Zij1z9bFiHR
* Optimise parallel as.sir(): row-batch mode when n_cols < n_cores
Previously parallel dispatch only parallelised by column, so a 6-column
dataset on a 16-core machine used at most 6 cores with the other 10 idle.
For large n this also caused memory-bandwidth saturation (each worker did
a full n-row scan of clinical_breakpoints simultaneously).
New row-batch mode (fork path, R >= 4.0, non-Windows):
pieces_per_col = ceil(n_cores / n_cols)
Jobs = n_cols × pieces_per_col (≈ n_cores jobs total)
Each job: one column × one row slice
Benefits:
- All cores stay busy regardless of column count
- Per-worker memory footprint shrinks by pieces_per_col ×
- Breakpoints lookup cache pressure reduced per worker
PSOCK path (Windows / R < 4.0) is unchanged: per-job serialisation
overhead makes row batching unprofitable there.
run_as_sir_column() gains an optional `rows` parameter (NULL = all rows,
backward-compatible). Results are reassembled via as.sir(c(as.character(.)))
which is safe for already-clean SIR values.
https://claude.ai/code/session_012DXCXbZUC54Zij1z9bFiHR
* Fix info=FALSE ignored when no breakpoints found in as_sir_method
Operator-precedence bug at line 1601:
if (isTRUE(info) && nrow(df_unique) < 10 || nrow(breakpoints) == 0)
R evaluates && before ||, so this was equivalent to:
(isTRUE(info) && nrow(df_unique) < 10) || (nrow(breakpoints) == 0)
When nrow(breakpoints) == 0 (e.g. cefoxitin / flucloxacillin / mupirocin
against E. coli in EUCAST) the intro message was always printed regardless
of info. Fix: add parentheses so info gates both conditions:
isTRUE(info) && (nrow(df_unique) < 10 || nrow(breakpoints) == 0)
Also pass print = isTRUE(info) to progress_ticker so the progress bar
(which prints intro_txt as its title) is suppressed when info = FALSE.
https://claude.ai/code/session_012DXCXbZUC54Zij1z9bFiHR
* Fix cli formatting in as.sir() messages
- stop_if for empty ab_cols: wrap as.mic() and as.disk() in
{.help [{.fun ...}](...)} for clickable links in cli output
- Parallel mode message: use {.field col} formatting for column names
and quotes = FALSE in vector_and(), consistent with the rest of the
codebase (avoids double-quoting from both font_bold and quotes="'")
https://claude.ai/code/session_012DXCXbZUC54Zij1z9bFiHR
* Use font_bold() inside {.field} for column names in parallel message
Convention: paste0("{.field ", font_bold(col), "}") gives bold green
column names without quotation marks, consistent with the rest of the
codebase (e.g. the 'Cleaning values' message in run_as_sir_column).
https://claude.ai/code/session_012DXCXbZUC54Zij1z9bFiHR
* Add collapse = NULL to font_bold() for column name vectors
font_bold() without collapse = NULL joins a vector with "" into a single
string, breaking paste0() element-wise formatting for length > 1 vectors.
https://claude.ai/code/session_012DXCXbZUC54Zij1z9bFiHR
* Add tools/ to .Rbuildignore
Keeps the benchmark script out of the built package tarball.
https://claude.ai/code/session_012DXCXbZUC54Zij1z9bFiHR
---------
Co-authored-by: Claude <noreply@anthropic.com>
* fix: convert Python lists to R vectors in wrapper to prevent R list coercion errors
Fixes#267. Python lists passed to R functions via rpy2 are received as
R lists, not R character/numeric vectors. This causes is.mic(), is.sir(),
is.disk() etc. to return length > 1 logicals, which break R's && operator.
Added convert_to_r() helper that maps Python list/tuple to the appropriate
typed R vector (StrVector, IntVector, FloatVector) based on element types.
The r_to_python decorator now applies this to all args and kwargs before
calling the R function.
* docs: instruct Claude to install git and gh before computing version
