kristjan/cjvt-valency
1
0
Fork 1
Code Issues Pull Requests Releases Wiki Activity

filtering some words in index

Browse Source
This commit is contained in:
voje
2019-04-27 20:24:11 +02:00
parent fd94627fdb
commit 81395890ab
16 changed files with 398 additions and 10 deletions
Download Patch File Download Diff File Expand all files Collapse all files

13
Makefile
View File

@@ -11,9 +11,14 @@ MAKE_ROOT = $(shell pwd)
# SSJ_FILE = "$(MAKE_ROOT)/data/samples/ssj_xml/ssj500k-sl.body.sample.xml"
SSJ_FILE = "$(MAKE_ROOT)/data/ssj_file_link"
# KRES_FOLDER = "$(MAKE_ROOT)/data/samples/kres_xml"
KRES_FOLDER = "$(MAKE_ROOT)/data/kres_xml_folder_link"
# KRES_FOLDER = "$(MAKE_ROOT)/data/kres_xml_folder_link"
KRES_FOLDER = "/home/kristjan/kres_data/payload/kres_xml"
# KRES_SRL_FOLDER = "$(MAKE_ROOT)/data/samples/kres_srl_json"
KRES_SRL_FOLDER = "$(MAKE_ROOT)/data/kres_json_folder_link"
# KRES_SRL_FOLDER = "$(MAKE_ROOT)/data/kres_json_folder_link"
KRES_SRL_FOLDER = "/home/kristjan/kres_data/payload/kres_json"
# This file comes with the source code. Make sure you unpack it and name it right.
SSKJ_WORDLIST = "$(MAKE_ROOT)/data/wordlist.json"
OUTPUT = "db"
# OUTPUT = "file"
@@ -24,7 +29,7 @@ DBADDR = "0.0.0.0:27017" # don't use localhost
# create it from env.default
include env.local
N_CORES = 5
N_CORES = 3
# insert kres files into database in chunks, for fewer connections
KRES_CHUNK_SIZE = 30
@@ -59,6 +64,7 @@ python-env:
python-env-install:
pip3 install -e src/pkg/cjvt-corpusparser/.
pip3 install -e src/pkg/valency/.
pip3 install -e src/pkg/seqparser/.
# from inside python-env container:
data/samples:
@@ -103,6 +109,7 @@ backend-prepare-db:
cd ./src/backend_flask; python3 app.py \
--config-file ./conf_files/dev_conf.yaml \
--dbuser $(DB_USR_USER) --dbpass $(DB_USR_PASS) --dbaddr $(DBADDR) \
--sskj-wordlist $(SSKJ_WORDLIST) \
--prepare-db
backend-dev: