Public/kotaemon
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Put the preparation step in FileIndex to on_start (#20)

Browse Source
This commit is contained in:
Duc Nguyen (john) 2024-04-10 19:30:45 +07:00 committed by GitHub
parent b507eef541
commit f3e82b2e70
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
2 changed files with 21 additions and 17 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

16
libs/ktem/ktem/index/file/index.py
View File

@ -81,11 +81,6 @@ class FileIndex(BaseIndex):
self._index_ui_cls: Type self._index_ui_cls: Type
self._index_ui: Any = None self._index_ui: Any = None
self._setup_indexing_cls()
self._setup_retriever_cls()
self._setup_file_index_ui_cls()
self._setup_file_selector_ui_cls()
self._default_settings: dict[str, dict] = {} self._default_settings: dict[str, dict] = {}
self._setting_mappings: dict[str, dict] = {} self._setting_mappings: dict[str, dict] = {}
@ -249,10 +244,6 @@ class FileIndex(BaseIndex):
# user's modification # user's modification
config.update(self.config) config.update(self.config)
# clean
file_types_str = config["supported_file_types"]
file_types = [each.strip() for each in file_types_str.split(",")]
config["supported_file_types"] = file_types
self.config = config self.config = config
# create the resources # create the resources
@ -270,6 +261,13 @@ class FileIndex(BaseIndex):
self._docstore.drop() self._docstore.drop()
shutil.rmtree(self._fs_path) shutil.rmtree(self._fs_path)
def on_start(self):
"""Setup the classes and hooks"""
self._setup_indexing_cls()
self._setup_retriever_cls()
self._setup_file_index_ui_cls()
self._setup_file_selector_ui_cls()
def get_selector_component_ui(self): def get_selector_component_ui(self):
if self._selector_ui is None: if self._selector_ui is None:
self._selector_ui = self._selector_ui_cls(self._app, self) self._selector_ui = self._selector_ui_cls(self._app, self)

22
libs/ktem/ktem/index/file/ui.py
View File

@ -42,14 +42,17 @@ class DirectoryUpload(BasePage):
def __init__(self, app, index): def __init__(self, app, index):
super().__init__(app) super().__init__(app)
self._index = index self._index = index
self._supported_file_types = self._index.config.get("supported_file_types", []) self._supported_file_types_str = self._index.config.get(
"supported_file_types", ""
)
self._supported_file_types = [
each.strip() for each in self._supported_file_types_str.split(",")
]
self.on_building_ui() self.on_building_ui()
def on_building_ui(self): def on_building_ui(self):
with gr.Accordion(label="Directory upload", open=False): with gr.Accordion(label="Directory upload", open=False):
gr.Markdown( gr.Markdown(f"Supported file types: {self._supported_file_types_str}")
f"Supported file types: {', '.join(self._supported_file_types)}",
)
self.path = gr.Textbox( self.path = gr.Textbox(
placeholder="Directory path...", lines=1, max_lines=1, container=False placeholder="Directory path...", lines=1, max_lines=1, container=False
) )
@ -69,7 +72,12 @@ class FileIndexPage(BasePage):
def __init__(self, app, index): def __init__(self, app, index):
super().__init__(app) super().__init__(app)
self._index = index self._index = index
self._supported_file_types = self._index.config.get("supported_file_types", []) self._supported_file_types_str = self._index.config.get(
"supported_file_types", ""
)
self._supported_file_types = [
each.strip() for each in self._supported_file_types_str.split(",")
]
self.selected_panel_false = "Selected file: (please select above)" self.selected_panel_false = "Selected file: (please select above)"
self.selected_panel_true = "Selected file: {name}" self.selected_panel_true = "Selected file: {name}"
# TODO: on_building_ui is not correctly named if it's always called in # TODO: on_building_ui is not correctly named if it's always called in
@ -80,9 +88,7 @@ class FileIndexPage(BasePage):
def upload_instruction(self) -> str: def upload_instruction(self) -> str:
msgs = [] msgs = []
if self._supported_file_types: if self._supported_file_types:
msgs.append( msgs.append(f"- Supported file types: {self._supported_file_types_str}")
f"- Supported file types: {', '.join(self._supported_file_types)}"
)
if max_file_size := self._index.config.get("max_file_size", 0): if max_file_size := self._index.config.get("max_file_size", 0):
msgs.append(f"- Maximum file size: {max_file_size} MB") msgs.append(f"- Maximum file size: {max_file_size} MB")