fix(dependencies): upgrade hyfi to 1.2.14

entelecheia · Jul 18, 2023 · ca9ac2b · ca9ac2b
1 parent 1536430
commit ca9ac2b
Show file tree

Hide file tree

Showing 18 changed files with 28 additions and 104 deletions.
diff --git a/poetry.lock b/poetry.lock
diff --git a/pyproject.toml b/pyproject.toml
@@ -14,7 +14,7 @@ corprep = 'corprep.__cli__:main'
 
 [tool.poetry.dependencies]
 python = ">=3.8.1,<3.12"
-hyfi = "^1.2.13"
+hyfi = "^1.2.14"
 ekonlpy = "^1.1.7"
 
 [tool.poetry.group.dev]

diff --git a/src/corprep/conf/cmd/copy_conf.yaml b/src/corprep/conf/cmd/copy_conf.yaml
@@ -1,6 +1,6 @@
 # @package _global_
 defaults:
-  - /copier: conf
+  - /copier: __init__
 
 _target_: hyfi.__cli__.run_copy
 copier:

diff --git a/src/corprep/conf/copier/conf.yaml → src/corprep/conf/copier/__init__.yaml b/src/corprep/conf/copier/conf.yaml → src/corprep/conf/copier/__init__.yaml
@@ -9,7 +9,9 @@ filetypes:
   - py
 # User-chosen additional file exclusion patterns.
 exclude:
-# User-chosen additional file skip patterns.
+# If `True`, exclude files that match the patterns for testing. (default: True)
+exclude_test_files: true
+# Skip files that already exist?
 skip_if_exists: false
 # Delete `dst_path` if there's an error?
 cleanup_on_error: false

diff --git a/src/corprep/conf/pipe/__init__.yaml b/src/corprep/conf/pipe/__init__.yaml
@@ -4,7 +4,7 @@ run_with:
 name:
 desc:
 env:
-use_pipe_obj: true
-pipe_obj_arg_name:
-return_pipe_obj: false
+use_pipe_obj: true # if true, the pipe target function will be called with the pipe object as the first argument
+pipe_obj_arg_name: # if use_pipe_obj is true, the pipe object will be passed to the pipe target function with this argument name
+return_pipe_obj: false # if true, the pipe target function will return the pipe object instead of the return value
 verbose: false
diff --git a/src/corprep/conf/pipe/dataset_filter.yaml b/src/corprep/conf/pipe/dataset_filter.yaml
@@ -0,0 +1,13 @@
+defaults:
+  - __general_external_funcs__
+
+run: corprep.datasets.preprocessing.filter.filter_dataset
+run_with:
+  num_proc: 1
+  batched: true
+  text_col: bodyText
+  verbose: false
+use_pipe_obj: true
+return_pipe_obj: false
+# 은는이가도과와에의등
+# [ ()\'",·]
diff --git a/src/corprep/conf/pipe/load_raw_dataset.yaml → src/corprep/conf/pipe/dataset_load_raw.yaml b/src/corprep/conf/pipe/load_raw_dataset.yaml → src/corprep/conf/pipe/dataset_load_raw.yaml
diff --git a/src/corprep/conf/pipe/save_dataset.yaml → src/corprep/conf/pipe/dataset_save.yaml b/src/corprep/conf/pipe/save_dataset.yaml → src/corprep/conf/pipe/dataset_save.yaml
diff --git a/src/corprep/conf/pipe/tokenize_dataset.yaml → src/corprep/conf/pipe/dataset_tokenize.yaml b/src/corprep/conf/pipe/tokenize_dataset.yaml → src/corprep/conf/pipe/dataset_tokenize.yaml
diff --git a/src/corprep/conf/pipe/load_dataframes.yaml b/src/corprep/conf/pipe/load_dataframes.yaml
diff --git a/src/corprep/conf/pipe/save_dataframes.yaml b/src/corprep/conf/pipe/save_dataframes.yaml
diff --git a/src/corprep/conf/pipe/test_preprocessing.yaml b/src/corprep/conf/pipe/test_preprocessing.yaml
diff --git a/src/corprep/conf/pipeline/__test_dataframe__.yaml b/src/corprep/conf/pipeline/__test_dataframe__.yaml
diff --git a/src/corprep/conf/pipeline/__test_general__.yaml b/src/corprep/conf/pipeline/__test_general__.yaml
diff --git a/src/corprep/conf/pipeline/datasets.yaml b/src/corprep/conf/pipeline/datasets.yaml
@@ -1,8 +1,8 @@
 defaults:
   - __init__
-  - /pipe@pipe1: load_raw_dataset
-  - /pipe@pipe2: tokenize_dataset
-  - /pipe@pipe3: save_dataset
+  - /pipe@pipe1: dataset_load_raw
+  - /pipe@pipe2: dataset_tokenize
+  - /pipe@pipe3: dataset_save
 
 use_task_as_initial_object: true
 steps:

diff --git a/src/corprep/conf/project/__test__.yaml b/src/corprep/conf/project/__test__.yaml
diff --git a/src/corprep/conf/task/__test__.yaml b/src/corprep/conf/task/__test__.yaml
diff --git a/src/corprep/conf/workflow/__test__.yaml b/src/corprep/conf/workflow/__test__.yaml