From 58f3a2e34a18a0f004922f203ccece7305df0c08 Mon Sep 17 00:00:00 2001
From: Yoann Schneider <yschneider@teklia.com>
Date: Wed, 16 Nov 2022 15:04:06 +0000
Subject: [PATCH] fix tokens retrieval
---
dan/datasets/extract/extract_from_arkindex.py | 12 +++++-------
1 file changed, 5 insertions(+), 7 deletions(-)
diff --git a/dan/datasets/extract/extract_from_arkindex.py b/dan/datasets/extract/extract_from_arkindex.py
index b0274f98..ea7860de 100644
--- a/dan/datasets/extract/extract_from_arkindex.py
+++ b/dan/datasets/extract/extract_from_arkindex.py
@@ -213,9 +213,11 @@ class ArkindexExtractor:
count = 0
for entity in entities["results"]:
- start_token, end_token = self.tokens[
- entity["entity"]["metas"]["subtype"]
- ]
+ matching_tokens = self.tokens[entity["entity"]["metas"]["subtype"]]
+ start_token, end_token = (
+ matching_tokens["start"],
+ matching_tokens["end"],
+ )
text, count = insert_token(
text,
count,
@@ -349,10 +351,6 @@ def run(
if load_entities:
assert tokens, "Please provide the entities to match."
- # Get and initialize the parameters.
- os.makedirs(IMAGES_DIR, exist_ok=True)
- os.makedirs(LABELS_DIR, exist_ok=True)
-
# Login to arkindex.
assert (
"ARKINDEX_API_URL" in os.environ
--
GitLab