FastaLake parameter presets (v5).
Legacy entrapment observations refer to the historical 592-sample MicrobPredict sweep;
they do not calibrate the current whole workflow.
Use via CLI:
fasta-lake infer --preset stringent ...
fasta-lake export --preset default ...
Or programmatically:
from fasta_lake.presets import apply_preset
cfg = apply_preset(cfg, "stringent")
dataclass
Immutable preset definition — a named bundle of parameter values.
Source code in fasta_lake/presets.py
| @dataclass(frozen=True)
class Preset:
"""Immutable preset definition — a named bundle of parameter values."""
name: str
description: str
# Stage 1-3 (de novo peptide filter)
denovo_min_length: int
denovo_max_length: int = 50
min_score: float = 0.0
top_percent: float = 0.30
# Stage 4 (SAGE search)
search_min_length: int = 9
search_max_length: int = 50
search_missed_cleavages: int = 2
search_max_variable_mods: int = 2
search_precursor_ppm: float = 10.0
search_fdr: float = 0.01
# Pipeline features
kmer_rescue: bool = False
cluster_enabled: bool = True
cluster_identity: float = 0.97
infer_strategy: str = "species_budget"
infer_min_peptides: int = 2
def as_dict(self) -> dict:
"""Return a dataclass dictionary of this preset's settings."""
from dataclasses import asdict
return asdict(self)
|
Return a dataclass dictionary of this preset's settings.
Source code in fasta_lake/presets.py
| def as_dict(self) -> dict:
"""Return a dataclass dictionary of this preset's settings."""
from dataclasses import asdict
return asdict(self)
|
get_preset(name: str) -> Preset
Return the named preset or raise ValueError listing the available names.
Source code in fasta_lake/presets.py
| def get_preset(name: str) -> Preset:
"""Return the named preset or raise ValueError listing the available names."""
if name not in PRESETS:
choices = ", ".join(sorted(PRESETS.keys()))
raise ValueError(f"Unknown preset '{name}'. Choose from: {choices}")
return PRESETS[name]
|
apply_preset(cfg, preset_name: str)
Apply preset fields to an existing RunConfig-like object.
Source code in fasta_lake/presets.py
| def apply_preset(cfg, preset_name: str):
"""Apply preset fields to an existing RunConfig-like object."""
preset = get_preset(preset_name)
for field_name, value in preset.as_dict().items():
if field_name in ("name", "description"):
continue
if hasattr(cfg, field_name):
setattr(cfg, field_name, value)
return cfg
|
preset_sage_config(base_config: dict, preset_name: str) -> dict
Apply preset to a SAGE JSON config dict. Returns modified copy.
Source code in fasta_lake/presets.py
| def preset_sage_config(base_config: dict, preset_name: str) -> dict:
"""Apply preset to a SAGE JSON config dict. Returns modified copy."""
import json
preset = get_preset(preset_name)
cfg = json.loads(json.dumps(base_config)) # deep copy
cfg["database"]["enzyme"]["min_len"] = preset.search_min_length
cfg["database"]["enzyme"]["max_len"] = preset.search_max_length
cfg["database"]["enzyme"]["missed_cleavages"] = preset.search_missed_cleavages
cfg["database"]["max_variable_mods"] = preset.search_max_variable_mods
cfg["precursor_tol"] = {"ppm": [-preset.search_precursor_ppm, preset.search_precursor_ppm]}
return cfg
|
describe_preset(name: str) -> str
Human-readable description of what a preset does.
Source code in fasta_lake/presets.py
| def describe_preset(name: str) -> str:
"""Human-readable description of what a preset does."""
p = get_preset(name)
lines = [
f"PRESET: {p.name.upper()}",
p.description,
"",
" Stage 1-3 (de novo evidence filter):",
f" min_length : {p.denovo_min_length}",
f" max_length : {p.denovo_max_length}",
f" top_percent : {p.top_percent}",
f" min_score : {p.min_score}",
"",
" Stage 4 (SAGE search):",
f" min_length : {p.search_min_length}",
f" max_length : {p.search_max_length}",
f" missed_cleavages: {p.search_missed_cleavages}",
f" max_var_mods : {p.search_max_variable_mods}",
f" precursor_ppm : ±{p.search_precursor_ppm}",
f" q_threshold : {p.search_fdr}",
"",
" Pipeline:",
f" clustering : {p.cluster_identity * 100:.0f}% identity",
f" inference : {p.infer_strategy} (min {p.infer_min_peptides} peptides)",
f" kmer_rescue : {p.kmer_rescue}",
]
return "\n".join(lines)
|