From d6f72188afcd4d90b556bd339e6054b80be3d190 Mon Sep 17 00:00:00 2001
From: "louise.deleger" <louise.deleger@inra.fr>
Date: Wed, 7 Apr 2021 16:09:11 +0200
Subject: [PATCH 1/4] additional blacklisted terms
---
ancillaries/blacklist.txt | 16 ++++++++++++----
1 file changed, 12 insertions(+), 4 deletions(-)
diff --git a/ancillaries/blacklist.txt b/ancillaries/blacklist.txt
index c505a5ce..fca79323 100644
--- a/ancillaries/blacklist.txt
+++ b/ancillaries/blacklist.txt
@@ -19,7 +19,7 @@ flagellum
harvesting antenna
host
industry
-IV secretion system
+iv secretion system
light - harvesting antenna
medium
membrane
@@ -31,8 +31,8 @@ single polar flagellum
spore
supply
tool
-type III secretion system
-type IV secretion system
+type iii secretion system
+type iv secretion system
vector
antenna
clinical
@@ -130,4 +130,12 @@ granular
community
communities
well
-wells
\ No newline at end of file
+wells
+phylogenetic tree
+coalescent tree
+isomorphic tree
+phyletic tree
+gene tree
+eukaryotic tree
+tree of life
+maximum-likelihood tree
--
GitLab
From 07c225a1ca9d7f77f1894408d159117c88cb40d6 Mon Sep 17 00:00:00 2001
From: "louise.deleger" <louise.deleger@inra.fr>
Date: Wed, 7 Apr 2021 16:09:51 +0200
Subject: [PATCH 2/4] lower case for blacklist
---
plans/tomap-habitats-generic.plan | 4 ++--
plans/tomap-habitats.plan | 2 +-
2 files changed, 3 insertions(+), 3 deletions(-)
diff --git a/plans/tomap-habitats-generic.plan b/plans/tomap-habitats-generic.plan
index 7ca9a1fd..93683ca0 100644
--- a/plans/tomap-habitats-generic.plan
+++ b/plans/tomap-habitats-generic.plan
@@ -253,7 +253,7 @@
<!-- Remove terms with blacklisted syntactic heads -->
<delete-blacklisted-heads class="Action">
- <target>documents.sections.layer:habitats[@explain_significant-head in "ancillaries/blacklisted-heads.txt" and not (@concept-name =~ " " and (str:lower(@lemma-string) ?= str:lower(@concept-name) or @score == "1.0"))]</target>
+ <target>documents.sections.layer:habitats[str:lower(@explain_significant-head) in "ancillaries/blacklisted-heads.txt" and not (@concept-name =~ " " and (str:lower(@lemma-string) ?= str:lower(@concept-name) or @score == "1.0"))]</target>
<action>remove:habitats</action>
<removeFromLayer/>
</delete-blacklisted-heads>
@@ -449,7 +449,7 @@
<!-- remove blacklisted terms -->
<delete-blacklisted class="Action">
- <target>documents.sections.layer:habitats[@selected == "true" and @lemma-string in "ancillaries/blacklist.txt"]</target>
+ <target>documents.sections.layer:habitats[@selected == "true" and str:lower(@lemma-string) in "ancillaries/blacklist.txt"]</target>
<action>remove:habitats</action>
<removeFromLayer/>
</delete-blacklisted>
diff --git a/plans/tomap-habitats.plan b/plans/tomap-habitats.plan
index e5dcec1c..f0f5f18e 100644
--- a/plans/tomap-habitats.plan
+++ b/plans/tomap-habitats.plan
@@ -449,7 +449,7 @@
<!-- remove blacklisted terms -->
<delete-blacklisted class="Action">
- <target>documents.sections.layer:habitats[@selected == "true" and @lemma-string in "ancillaries/blacklist.txt"]</target>
+ <target>documents.sections.layer:habitats[@selected == "true" and str:lower(@lemma-string) in "ancillaries/blacklist.txt"]</target>
<action>remove:habitats</action>
<removeFromLayer/>
</delete-blacklisted>
--
GitLab
From 595a2e2b7633cb90d46d252990a9477012245f32 Mon Sep 17 00:00:00 2001
From: "louise.deleger" <louise.deleger@inra.fr>
Date: Wed, 7 Apr 2021 16:12:51 +0200
Subject: [PATCH 3/4] changed PUBMED_BATCHES_HOME
---
config/config.yaml | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/config/config.yaml b/config/config.yaml
index 9b2c74a1..b416bf10 100644
--- a/config/config.yaml
+++ b/config/config.yaml
@@ -34,7 +34,7 @@ DSMZ_RESULT : "corpora/dsmz/test-3.3.txt"
## pubmed
PUBMED_FOLDER_NAME : "microbes-2019"
PUBMED_CORPUS_HOME : "corpora/microbes-2019"
-PUBMED_BATCHES_HOME : "corpora/microbes-2019/batch"
+PUBMED_BATCHES_HOME : "corpora/pubmed/batches"
PUBMED_HABITAT_RESULT : "ancillaries/Florilege/2019-12-12/PubMed-Habitat-2019-12-12.txt"
PUBMED_PHENOTYPE_RESULT : "ancillaries/Florilege/2019-12-12/PubMed-Phenotype-2019-12-12.txt"
--
GitLab
From 3d127d0a268d354823d8582a695cc367e6e65e6d Mon Sep 17 00:00:00 2001
From: Mouhamadou Ba <mouhamadou.ba@inra.fr>
Date: Wed, 7 Apr 2021 17:04:36 +0200
Subject: [PATCH 4/4] Revert "changed PUBMED_BATCHES_HOME"
This reverts commit 595a2e2b7633cb90d46d252990a9477012245f32.
---
config/config.yaml | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/config/config.yaml b/config/config.yaml
index b416bf10..9b2c74a1 100644
--- a/config/config.yaml
+++ b/config/config.yaml
@@ -34,7 +34,7 @@ DSMZ_RESULT : "corpora/dsmz/test-3.3.txt"
## pubmed
PUBMED_FOLDER_NAME : "microbes-2019"
PUBMED_CORPUS_HOME : "corpora/microbes-2019"
-PUBMED_BATCHES_HOME : "corpora/pubmed/batches"
+PUBMED_BATCHES_HOME : "corpora/microbes-2019/batch"
PUBMED_HABITAT_RESULT : "ancillaries/Florilege/2019-12-12/PubMed-Habitat-2019-12-12.txt"
PUBMED_PHENOTYPE_RESULT : "ancillaries/Florilege/2019-12-12/PubMed-Phenotype-2019-12-12.txt"
--
GitLab