Skip to content

Commit 1ebf9ea

Browse files
committed
Swap out es model to bigger one
1 parent ad8a434 commit 1ebf9ea

File tree

2 files changed

+19
-17
lines changed

2 files changed

+19
-17
lines changed

Dockerfile

Lines changed: 18 additions & 16 deletions
Original file line numberDiff line numberDiff line change
@@ -1,4 +1,4 @@
1-
FROM ubuntu:20.04
1+
FROM ubuntu:22.04
22
ENV DEBIAN_FRONTEND noninteractive
33

44
LABEL org.opencontainers.image.title "FollowTheMoney File Ingestors"
@@ -129,21 +129,23 @@ RUN pip3 install --no-cache-dir --prefer-binary --upgrade setuptools wheel
129129
RUN pip3 install --no-cache-dir --no-binary "tesserocr" -r /tmp/requirements.txt
130130

131131
# Install spaCy models
132-
RUN python3 -m spacy download en_core_web_sm \
133-
&& python3 -m spacy download de_core_news_sm \
134-
&& python3 -m spacy download fr_core_news_sm \
135-
&& python3 -m spacy download es_core_news_sm
136-
RUN python3 -m spacy download ru_core_news_sm \
137-
&& python3 -m spacy download pt_core_news_sm \
138-
&& python3 -m spacy download ro_core_news_sm \
139-
&& python3 -m spacy download mk_core_news_sm
140-
RUN python3 -m spacy download el_core_news_sm \
141-
&& python3 -m spacy download pl_core_news_sm \
142-
&& python3 -m spacy download it_core_news_sm \
143-
&& python3 -m spacy download lt_core_news_sm \
144-
&& python3 -m spacy download nl_core_news_sm \
145-
&& python3 -m spacy download nb_core_news_sm \
146-
&& python3 -m spacy download da_core_news_sm
132+
RUN python3 -m spacy download en_core_web_sm
133+
RUN python3 -m spacy download es_core_news_lg
134+
RUN python3 -m spacy download de_core_news_sm
135+
RUN python3 -m spacy download fr_core_news_sm
136+
RUN python3 -m spacy download da_core_news_sm
137+
RUN python3 -m spacy download uk_core_news_sm
138+
RUN python3 -m spacy download ru_core_news_sm
139+
RUN python3 -m spacy download pt_core_news_sm
140+
RUN python3 -m spacy download ro_core_news_sm
141+
RUN python3 -m spacy download mk_core_news_sm
142+
RUN python3 -m spacy download el_core_news_sm
143+
RUN python3 -m spacy download pl_core_news_sm
144+
RUN python3 -m spacy download it_core_news_sm
145+
RUN python3 -m spacy download lt_core_news_sm
146+
RUN python3 -m spacy download nl_core_news_sm
147+
RUN python3 -m spacy download nb_core_news_sm
148+
RUN python3 -m spacy download sv_core_news_sm
147149
# RUN python3 -m spacy download zh_core_web_sm
148150

149151
COPY . /ingestors

ingestors/settings.py

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -23,7 +23,7 @@
2323
"eng": "en_core_web_sm",
2424
"deu": "de_core_news_sm",
2525
"fra": "fr_core_news_sm",
26-
"spa": "es_core_news_sm",
26+
"spa": "es_core_news_lg",
2727
"rus": "ru_core_news_sm",
2828
"por": "pt_core_news_sm",
2929
"ron": "ro_core_news_sm",

0 commit comments

Comments
 (0)