Revert back to model from v0.8.2

This reverts commit 23c77f2.
zellerlab · Aug 23, 2021 · 8554595 · 8554595
1 parent 23c77f2
commit 8554595
Show file tree

Hide file tree

Showing 10 changed files with 130 additions and 4,109 deletions.
diff --git a/gecco/cli/commands/annotate.py b/gecco/cli/commands/annotate.py
@@ -59,7 +59,7 @@ def doc(cls, fast=False):  # noqa: D102
             -e <e>, --e-filter <e>        the e-value cutoff for protein domains
                                           to be included.
             -p <p>, --p-filter <p>        the p-value cutoff for protein domains
-                                          to be included. [default: 1e-5]
+                                          to be included. [default: 1e-9]
 
         Parameters - Debug:
             --hmm <hmm>                   the path to one or more alternative

diff --git a/gecco/cli/commands/cv.py b/gecco/cli/commands/cv.py
@@ -53,7 +53,7 @@ def doc(cls, fast=False):  # noqa: D102
             -e <e>, --e-filter <e>          the e-value cutoff for domains to
                                             be included.
             -p <p>, --p-filter <p>          the p-value cutoff for domains
-                                            to be included. [default: 1e-5]
+                                            to be included. [default: 1e-9]
 
         Parameters - Training Data:
             --no-shuffle                    disable shuffling of the data

diff --git a/gecco/cli/commands/run.py b/gecco/cli/commands/run.py
@@ -68,7 +68,7 @@ def doc(cls, fast=False):  # noqa: D102
             -e <e>, --e-filter <e>        the e-value cutoff for protein domains
                                           to be included.
             -p <p>, --p-filter <p>        the p-value cutoff for protein domains
-                                          to be included. [default: 1e-5]
+                                          to be included. [default: 1e-9]
 
         Parameters - Cluster Detection:
             -c, --cds <N>                 the minimum number of coding sequences a

diff --git a/gecco/cli/commands/train.py b/gecco/cli/commands/train.py
@@ -54,7 +54,7 @@ def doc(cls, fast: bool = False) -> str:  # noqa: D102
             -e <e>, --e-filter <e>          the e-value cutoff for domains to
                                             be included.
             -p <p>, --p-filter <p>          the p-value cutoff for domains to
-                                            be included. [default: 1e-5]
+                                            be included. [default: 1e-9]
 
         Parameters - Training Data:
             --no-shuffle                    disable shuffling of the data

diff --git a/gecco/crf/model.pkl b/gecco/crf/model.pkl
diff --git a/gecco/crf/model.pkl.md5 b/gecco/crf/model.pkl.md5
@@ -1 +1 @@
-9415db745ec926e45d78c842736320f1
+39393c2a69c42b93305f9462cff1764a
diff --git a/gecco/hmmer/Pfam.ini b/gecco/hmmer/Pfam.ini
@@ -1,8 +1,7 @@
 [hmm]
 id = Pfam
 version = 34.0
-size = 6661
+size = 2683
 url = ftp://ftp.ebi.ac.uk/pub/databases/Pfam/releases/Pfam34.0/Pfam-A.hmm.gz
 relabel_with = s/(PF\d+).\d+/\1/
-md5 = c77f60de9076e01b7b1b2eb002ed3019
-
+md5 = e56170a43f91cc6afd41b696e045c107
diff --git a/gecco/interpro/interpro.json.gz b/gecco/interpro/interpro.json.gz
diff --git a/gecco/types/compositions.npz b/gecco/types/compositions.npz