From 64ffa8c0d1072b9f9928fd7419c6b384dcd0786d Mon Sep 17 00:00:00 2001
From: Tim O'Donnell <timodonnell@gmail.com>
Date: Fri, 16 Feb 2018 15:31:59 -0500
Subject: [PATCH] optimization
---
mhcflurry/class1_affinity_predictor.py | 49 ++++++++++++++++----------
mhcflurry/encodable_sequences.py | 2 ++
2 files changed, 33 insertions(+), 18 deletions(-)
diff --git a/mhcflurry/class1_affinity_predictor.py b/mhcflurry/class1_affinity_predictor.py
index ea4939b4..2ad081f6 100644
--- a/mhcflurry/class1_affinity_predictor.py
+++ b/mhcflurry/class1_affinity_predictor.py
@@ -848,20 +848,29 @@ class Class1AffinityPredictor(object):
(min_peptide_length, max_peptide_length) = (
self.supported_peptide_lengths)
- df["supported_peptide_length"] = (
- (df.peptide.str.len() >= min_peptide_length) &
- (df.peptide.str.len() <= max_peptide_length))
- if (~df.supported_peptide_length).any():
- msg = (
- "%d peptides have lengths outside of supported range [%d, %d]: "
- "%s" % (
- (~df.supported_peptide_length).sum(),
- min_peptide_length,
- max_peptide_length,
- str(df.ix[~df.supported_peptide_length].peptide.unique())))
- logging.warning(msg)
- if throw:
- raise ValueError(msg)
+
+ if (peptides.min_length < min_peptide_length or
+ peptides.max_length > max_peptide_length):
+ # Only compute this if needed
+ all_peptide_lengths_supported = False
+ df["supported_peptide_length"] = (
+ (df.sequence_length.len() >= min_peptide_length) &
+ (df.sequence_length.len() <= max_peptide_length))
+ if (~df.supported_peptide_length).any():
+ msg = (
+ "%d peptides have lengths outside of supported range [%d, %d]: "
+ "%s" % (
+ (~df.supported_peptide_length).sum(),
+ min_peptide_length,
+ max_peptide_length,
+ str(df.ix[~df.supported_peptide_length].peptide.unique())))
+ logging.warning(msg)
+ if throw:
+ raise ValueError(msg)
+ else:
+ # Handle common case efficiently.
+ df["supported_peptide_length"] = True
+ all_peptide_lengths_supported = True
if self.class1_pan_allele_models:
unsupported_alleles = [
@@ -904,12 +913,16 @@ class Class1AffinityPredictor(object):
logging.warning(msg)
if throw:
raise ValueError(msg)
+
for allele in query_alleles:
models = self.allele_to_allele_specific_models.get(allele, [])
- mask = (
- (df.normalized_allele == allele) &
- df.supported_peptide_length).values
- if mask.all():
+ if len(query_alleles) == 1 and all_peptide_lengths_supported:
+ mask = None
+ else:
+ mask = (
+ (df.normalized_allele == allele) &
+ df.supported_peptide_length).values
+ if mask is None or mask.all():
# Common case optimization
for (i, model) in enumerate(models):
df["model_single_%d" % i] = model.predict(peptides)
diff --git a/mhcflurry/encodable_sequences.py b/mhcflurry/encodable_sequences.py
index da4cf9dc..efe29c14 100644
--- a/mhcflurry/encodable_sequences.py
+++ b/mhcflurry/encodable_sequences.py
@@ -39,6 +39,8 @@ class EncodableSequences(object):
"sequence": numpy.array(sequences),
"sequence_length": numpy.array(sequences),
})
+ self.min_length = self.sequences_df.sequence_length.min()
+ self.max_length = self.sequences_df.sequence_length.max()
self.sequences = self.sequences_df.sequence.values
self.encoding_cache = {}
self.fixed_sequence_length = None
--
GitLab