MCPcopy Create free account

hub / github.com/blackrock/HOLA / functions

Functions1,098 in github.com/blackrock/HOLA

↓ 262 callersMethodpush
Append a new trial. Returns the assigned trial ID. This is an O(1) operation - no sorting is performed.
opt_engine/src/leaderboard.rs:333
↓ 125 callersMethoditer
Iterator over all trials in insertion order.
opt_engine/src/leaderboard.rs:465
↓ 112 callersMethodclone
(&self)
hola/src/hola_engine.rs:480
↓ 90 callersMethodask
Ask for the next trial to evaluate. Returns an error if `max_trials` has been reached.
hola/src/hola_engine.rs:1835
↓ 74 callersMethodask
Request the next trial to evaluate.
hola-py/src/lib.rs:671
↓ 68 callersMethodlen
(&self)
opt_engine/src/leaderboard.rs:418
↓ 65 callersMethodget
Get a trial by its ID. O(1): backed by an internal `trial_id -> index` map kept in sync with `trials` on every push, eviction, and after deserializat
opt_engine/src/leaderboard.rs:452
↓ 64 callersMethodtell
Tell the engine the result of a trial, returning the scored and ranked trial.
hola/src/hola_engine.rs:1879
↓ 56 callersMethodtell
Report the result of a trial. Returns the scored and ranked CompletedTrial.
hola-py/src/lib.rs:728
↓ 46 callersFunctionjson_request
( app: axum::Router, method: &str, uri: &str, body: Option<serde_json::Value>, )
hola/tests/integration/server.rs:122
↓ 44 callersMethodrun
( slf: Py<Self>, py: Python<'_>, func: &Bound<'_, PyAny>, n_trials: usize,
hola-py/src/lib.rs:1082
↓ 40 callersFunctionhttp_json
Make an HTTP request and return (status_code, json_body). Uses urllib only — no extra dependencies.
hola-py/tests/conftest.py:168
↓ 36 callersMethodis_empty
(&self)
opt_engine/src/leaderboard.rs:422
↓ 36 callersMethodtop_k
(&self, py: Python<'_>, k: usize, include_infeasible: bool)
hola-py/src/lib.rs:860
↓ 34 callersMethodtrial_count
Number of completed trials.
hola-py/src/lib.rs:978
↓ 28 callersMethodclone
(&self)
opt_engine/src/strategies/gmm.rs:783
↓ 24 callersFunctionminimal_config
()
hola/tests/integration/server.rs:28
↓ 23 callersMethodlen
(&self)
hola/src/hola_engine.rs:1082
↓ 22 callersMethodread
(&mut self, buf: &mut [u8])
hola-cli/src/main.rs:538
↓ 19 callersFunctioncreate_router
Create the Axum router for the engine server. Returns an error if any configured CORS origin is not a valid HTTP header value.
hola/src/server.rs:602
↓ 19 callersFunctionsingle_objective_config
(strategy_type: &str)
hola/src/hola_engine.rs:3147
↓ 15 callersMethodsuggest
(&self, space: &Self::Space)
opt_engine/src/strategies/sobol.rs:157
↓ 15 callersMethodtop_k
Get the top-k trials by rank.
hola/src/hola_engine.rs:2014
↓ 14 callersMethodas_str
(self)
hola/src/hola_engine.rs:969
↓ 13 callersFunction_register
(p: SingleObjectiveProblem)
hola-py/benchmarks/problems/single_objective.py:32
↓ 13 callersMethodadd_real
(mut self, name: &str, min: f64, max: f64)
hola/src/hola_engine.rs:270
↓ 13 callersMethodconnect
(url: &str, token: Option<String>)
hola-py/src/lib.rs:618
↓ 13 callersFunctiondominates
(a, b)
dashboard/app.js:367
↓ 12 callersFunctioncheck
(cond, msg)
dashboard/tests/xss_dom.test.mjs:43
↓ 12 callersFunctioncreate_router_with_options
Create the Axum router for the engine server with explicit server options. Returns an error if any configured CORS origin is not a valid HTTP header
hola/src/server.rs:610
↓ 12 callersMethodload
Load a full checkpoint, restoring both leaderboard and strategy state. This is the stable persistence API. Use `save` / `load` for checkpointing.
hola/src/hola_engine.rs:2183
↓ 12 callersMethodpareto_front
Compute the Pareto front (non-dominated feasible trials). Infeasible trials (those with any infinite objective) are excluded. Use `pareto_front_all()
opt_engine/src/leaderboard.rs:804
↓ 12 callersMethodupdate_objectives
Update objectives and re-scalarize (for mid-run dashboard adjustments). Persists the new objectives so that subsequent `tell()` calls use the updated
hola/src/hola_engine.rs:2126
↓ 12 callersFunctionvalid_config_for_validation
()
hola/tests/integration/hola_engine.rs:93
↓ 11 callersFunctionassert_config_error
(config: StudyConfig, expected: &[&str])
hola/tests/integration/hola_engine.rs:125
↓ 11 callersMethodpareto_front
( &self, py: Python<'_>, front: usize, include_infeasible: bool, )
hola-py/src/lib.rs:895
↓ 11 callersFunctionwrite_yaml_config
Write a YAML config file and return the path. Uses simple string formatting to avoid pyyaml dependency.
hola-py/tests/conftest.py:103
↓ 10 callersFunctioncheck_format_version_bytes
Cheaply validate the recorded format version from raw checkpoint bytes. On-disk checkpoints come in three shapes, two of which carry `metadata` at th
opt_engine/src/persistence.rs:332
↓ 10 callersMethodtop_k
Return the k trials with the lowest (best) observations. Infeasible trials (those with infinite observations) are excluded. Use `top_k_all()` to incl
opt_engine/src/leaderboard.rs:549
↓ 9 callersFunctionapiFetch
(url, options = {})
dashboard/app.js:68
↓ 9 callersMethodfrom_unit_cube
(&self, vec: &[f64])
opt_engine/src/spaces/combinators.rs:61
↓ 9 callersMethodsuggest
(&self, space: &Self::Space)
opt_engine/src/strategies/gmm.rs:802
↓ 8 callersFunction_register
(p: MultiObjectiveProblem)
hola-py/benchmarks/problems/multi_objective.py:29
↓ 8 callersFunction_wait_for_server
Poll GET /api/space until the server is ready.
hola-py/tests/conftest.py:198
↓ 8 callersMethodadd_integer
(mut self, name: &str, min: i64, max: i64)
hola/src/hola_engine.rs:300
↓ 8 callersFunctionauthorize_read
Read endpoints and the SSE stream require the token only when the server was started with read authentication enabled (`require_read_auth`).
hola/src/server.rs:230
↓ 8 callersMethodnon_dominated_sort
Perform NSGA-II non-dominated sorting on feasible trials. Returns trials grouped by rank (front). `fronts[0]` is the Pareto front (rank 1), `fronts[1
opt_engine/src/leaderboard.rs:1057
↓ 8 callersMethodsave_full_checkpoint
Save a full checkpoint (leaderboard + strategy state + config). The saved JSON has the format: ```json { "config": { ...StudyConfig... }, "checkpoint
hola/src/hola_engine.rs:2338
↓ 8 callersMethodsuggest
(&self, space: &Self::Space)
opt_engine/src/strategies/random.rs:108
↓ 8 callersMethodtrials
( &self, py: Python<'_>, sorted_by: &str, include_infeasible: bool, )
hola-py/src/lib.rs:936
↓ 8 callersMethodtrials
Get all trials with scoring and ranking.
hola/src/hola_engine.rs:2051
↓ 7 callersFunction_find_free_port
()
hola-py/tests/test_cli.py:31
↓ 7 callersFunction_unit_bounds
(n: int)
hola-py/benchmarks/problems/multi_objective.py:22
↓ 7 callersFunctionclearElement
(el)
dashboard/app.js:75
↓ 7 callersMethoddimensionality
(&self)
opt_engine/src/spaces/combinators.rs:51
↓ 7 callersMethodinverse
(&self, x: f64)
opt_engine/src/scales.rs:61
↓ 7 callersFunctionminMax
(arr)
dashboard/app.js:82
↓ 7 callersFunctionrenderAll
()
dashboard/app.js:463
↓ 7 callersFunctionscalarize_raw
Collapse multi-field raw metrics into a single scalar cost F(x). Implements the paper's formula: F(x) = Σ φ_i(f_i(x)) where φ_i(u) = P_i × (u − T_i)/
hola/src/hola_engine.rs:2532
↓ 7 callersMethodto_unit_cube
(&self, point: &f64)
opt_engine/src/spaces/continuous.rs:78
↓ 7 callersMethodto_unit_cube
(&self, point: &Self::Domain)
opt_engine/src/spaces/combinators.rs:55
↓ 6 callersFunctionatomic_write_json
Write JSON to a file atomically: write to a unique temp file, fsync, rename, then fsync the parent directory. Prevents data loss if the process crashe
opt_engine/src/persistence.rs:444
↓ 6 callersMethodcancel
Cancel a pending trial.
hola/src/hola_engine.rs:2001
↓ 6 callersFunctionjson_to_py
(py: Python<'_>, val: &serde_json::Value)
hola-py/src/lib.rs:1287
↓ 6 callersMethodload_full_checkpoint
Load a full checkpoint, restoring both leaderboard and strategy state. Handles both the new format (with `"config"` + `"checkpoint"` wrapper) and the
hola/src/hola_engine.rs:2401
↓ 6 callersMethodload_leaderboard_checkpoint
Load a leaderboard-only checkpoint, replacing the current trial history. The strategy is not restored; call with `refit()` afterward if using GMM.
hola/src/hola_engine.rs:2252
↓ 6 callersFunctionnormalize_server_url
Validate and normalize a server base URL. Requires an http/https scheme and a host, and strips any trailing slash so that paths joined as "{server}/a
hola-cli/src/main.rs:116
↓ 6 callersMethodsample_clamped
Sample from the GMM, clamped to [0, 1]^n. Non-finite drawn values (NaN/inf) are replaced with the cube center (0.5) before clamping, since `f64::clam
opt_engine/src/strategies/gmm.rs:366
↓ 6 callersFunctionwith_bearer_auth
( request: reqwest::RequestBuilder, token: Option<&str>, )
hola-cli/src/main.rs:140
↓ 6 callersMethodwith_max_size
Set an optional maximum number of stored trials (builder form). `None` (the default) leaves the leaderboard unbounded. `Some(0)` is normalized to unb
opt_engine/src/leaderboard.rs:213
↓ 5 callersFunction_assert_same_params
(actual, expected)
hola-py/tests/test_study_advanced.py:283
↓ 5 callersFunction_completed_trial
Run a single ask/tell and return the CompletedTrial for `metrics`.
hola-py/tests/test_ffi_roundtrip.py:29
↓ 5 callersFunction_extract_vec
(p: dict[str, float])
hola-py/benchmarks/functions/zdt.py:28
↓ 5 callersFunction_extract_vec
Extract ordered vector from dict with keys x0, x1, ... or x, x1, x2, ...
hola-py/benchmarks/functions/single_objective.py:27
↓ 5 callersMethodadd_categorical
(mut self, name: &str, choices: Vec<String>)
hola/src/hola_engine.rs:310
↓ 5 callersFunctionauthorize_mutation
Mutating endpoints always require the token when one is configured.
hola/src/server.rs:221
↓ 5 callersFunctionauto_strategy_test_config
(exploration_budget: usize, seed: u64)
hola/tests/integration/hola_engine.rs:1789
↓ 5 callersMethodclear
(&mut self)
opt_engine/src/leaderboard.rs:473
↓ 5 callersFunctioncount_priority_groups
Count the number of distinct priority groups in the objectives.
hola/src/hola_engine.rs:1566
↓ 5 callersFunctiondecide_exec_outcome
Decide what to do with a finished exec-mode command, independent of any network I/O so the decision can be unit-tested directly. A zero exit with val
hola-cli/src/main.rs:222
↓ 5 callersMethoddim
(&self)
opt_engine/src/strategies/gmm.rs:152
↓ 5 callersMethoddimensionality
(&self)
hola/src/hola_engine.rs:94
↓ 5 callersMethodfeasible_trials
Return only feasible trials (those with finite observations).
opt_engine/src/leaderboard.rs:528
↓ 5 callersFunctionfmtCell
(v)
dashboard/app.js:1109
↓ 5 callersMethodfrom_unit_cube
(&self, vec: &[f64])
opt_engine/src/spaces/continuous.rs:102
↓ 5 callersFunctionjson_request_with_headers
( app: axum::Router, method: &str, uri: &str, body: Option<serde_json::Value>, headers: &[
hola/tests/integration/server.rs:131
↓ 5 callersMethodlast
(&self)
opt_engine/src/leaderboard.rs:469
↓ 5 callersMethodload
(path: &str)
hola-py/src/lib.rs:657
↓ 5 callersMethodobjectives
Get the current objectives configuration.
hola/src/hola_engine.rs:2109
↓ 5 callersMethodquantile_threshold
Compute the quantile threshold for feasible observations. `quantile` should be in [0, 1]. Returns the observation value at that quantile. E.g., `quan
opt_engine/src/leaderboard.rs:660
↓ 5 callersFunctionread_capped_reader
Read at most `limit` bytes from a reader, erroring if the source yields more. The reader is bounded to `limit + 1` so reaching the extra byte signals
opt_engine/src/persistence.rs:379
↓ 5 callersFunctionread_checkpoint_capped
Read a checkpoint file into memory with a byte-size cap. Rejects files whose reported length exceeds [`MAX_CHECKPOINT_BYTES`] before reading them, an
opt_engine/src/persistence.rs:307
↓ 5 callersMethodsave_leaderboard_checkpoint_to
Save a leaderboard-only checkpoint (trial history, no strategy state). Uses atomic writes (write-to-temp + fsync + rename) to prevent data loss.
hola/src/hola_engine.rs:2228
↓ 5 callersFunctionsetDot
(state)
dashboard/app.js:455
↓ 5 callersMethodset_params
Update the GMM parameters.
opt_engine/src/strategies/gmm.rs:756
↓ 5 callersMethodstore_trial
Append a trial record, updating the id index and enforcing `max_size`. Index update is O(1); eviction (only when a cap is set and exceeded) removes o
opt_engine/src/leaderboard.rs:293
↓ 5 callersMethodsuggest
(&self, space: &DynSpace)
hola/src/hola_engine.rs:544
↓ 5 callersMethodto_json
(&self)
opt_engine/src/persistence.rs:190
↓ 5 callersFunctionvectorize_raw
Compute per-group cost vector C(x) from raw metrics. Groups objectives by their explicit group label. Within each group, the group cost is: C_g(x) =
hola/src/hola_engine.rs:2554
↓ 4 callersFunction_extract_vec
(p: dict[str, float])
hola-py/benchmarks/functions/dtlz.py:28
next →1–100 of 1,098, ranked by callers