2023-08-11 22:02:51 +03:00
|
|
|
#!/usr/bin/env bash
|
|
|
|
|
set -e
|
|
|
|
|
|
|
|
|
|
# Description: This test checks if the number of bases and tables processed are as expected.
|
|
|
|
|
# Each base shows up as a directory in the output folder, hence check-num-dirs-output.sh
|
|
|
|
|
# Each table shows up as a file in a base directory, hence check-num-files-output.sh
|
|
|
|
|
|
|
|
|
|
SCRIPT_DIR=$(dirname "$(realpath "$0")")
|
|
|
|
|
cd "$SCRIPT_DIR"/.. || exit 1
|
|
|
|
|
|
|
|
|
|
OUTPUT_FOLDER_NAME=airtable-large
|
|
|
|
|
OUTPUT_DIR=$SCRIPT_DIR/structured-output/$OUTPUT_FOLDER_NAME
|
|
|
|
|
DOWNLOAD_DIR=$SCRIPT_DIR/download/$OUTPUT_FOLDER_NAME
|
2023-09-28 21:48:19 -07:00
|
|
|
max_processes=${MAX_PROCESSES:=$(python3 -c "import os; print(os.cpu_count())")}
|
2023-10-02 16:47:24 -04:00
|
|
|
CI=${CI:-"false"}
|
2023-08-11 22:02:51 +03:00
|
|
|
|
2023-09-21 14:51:08 -04:00
|
|
|
# shellcheck disable=SC1091
|
|
|
|
|
source "$SCRIPT_DIR"/cleanup.sh
|
2023-10-02 16:47:24 -04:00
|
|
|
function cleanup() {
|
|
|
|
|
cleanup_dir "$OUTPUT_DIR"
|
|
|
|
|
if [ "$CI" == "true" ]; then
|
|
|
|
|
cleanup_dir "$DOWNLOAD_DIR"
|
|
|
|
|
fi
|
|
|
|
|
}
|
|
|
|
|
trap cleanup EXIT
|
2023-09-21 14:51:08 -04:00
|
|
|
|
2023-08-11 22:02:51 +03:00
|
|
|
if [ -z "$AIRTABLE_PERSONAL_ACCESS_TOKEN" ]; then
|
|
|
|
|
echo "Skipping Airtable ingest test because the AIRTABLE_PERSONAL_ACCESS_TOKEN is not set."
|
|
|
|
|
exit 0
|
|
|
|
|
fi
|
|
|
|
|
|
|
|
|
|
# Provides component IDs such as LARGE_TEST_LIST_OF_PATHS,
|
|
|
|
|
# LARGE_TABLE_BASE_ID, LARGE_TABLE_TABLE_ID, and LARGE_BASE_BASE_ID
|
|
|
|
|
# shellcheck disable=SC1091
|
|
|
|
|
source ./scripts/airtable-test-helpers/component_ids.sh
|
|
|
|
|
|
|
|
|
|
PYTHONPATH=. ./unstructured/ingest/main.py \
|
|
|
|
|
airtable \
|
|
|
|
|
--download-dir "$DOWNLOAD_DIR" \
|
|
|
|
|
--personal-access-token "$AIRTABLE_PERSONAL_ACCESS_TOKEN" \
|
|
|
|
|
--list-of-paths "$LARGE_TEST_LIST_OF_PATHS" \
|
Feat: Create a naive hierarchy for elements (#1268)
## **Summary**
By adding hierarchy to unstructured elements, users will have more
information for implementing vector db/LLM chunking strategies. For
example, text elements could be queried by their preceding title
element. The hierarchy is implemented by a parent_id tag in the
element's metadata.
### Features
- Introduces a parent_id to ElementMetadata (The id of the parent
element, not a pointer)
- Creates a rule set for assigning hierarchies. Sensible default is
assigned, with an optional override parameter
- Sets element parent ids if there isn't an existing parent id or
matches the ruleset
### How it works
Hierarchies are assigned via a parent id field in element metadata.
Elements are read sequentially and evaluated against a ruleset. For
example take the following elements:
1. Title, "This is the Title"
2. Text, "this is the text"
And the ruleset: `{"title": ["text"]}`. When evaluated, the parent_id of
2 will be the id of 1. The algorithm for determining this is more
complex and resolves several edge cases, so please read the code for
further details.
### Schema Changes
```
@dataclass
class ElementMetadata:
coordinates: Optional[CoordinatesMetadata] = None
data_source: Optional[DataSourceMetadata] = None
filename: Optional[str] = None
file_directory: Optional[str] = None
last_modified: Optional[str] = None
filetype: Optional[str] = None
attached_to_filename: Optional[str] = None
+ parent_id: Optional[Union[str, uuid.UUID, NoID, UUID]] = None
+ category_depth: Optional[int] = None
...
```
### Testing
```
from unstructured.partition.auto import partition
from typing import List
elements = partition(filename="./unstructured/example-docs/fake-html.html", strategy="auto")
for element in elements:
print(
f"Category: {getattr(element, 'category', '')}\n"\
f"Text: {getattr(element, 'text', '')}\n"
f"ID: {element.id}\n" \
f"Parent ID: {element.metadata.parent_id}\n"\
f"Depth: {element.metadata.category_depth}\n" \
)
```
### Additional Notes
Implementing this feature revealed a possibly undesired side-effect in
how element metadata are processed. In
`unstructured/partition/common.py` the `_add_element_metadata` is
invoked as part of the `add_metadata_with_filetype` decorator for
filetype partitioning. This method is intended to add additional
information to the metadata generated with the element including
filename and filetype, however the existing metadata is merged into a
newly created metadata object rather than the other way around. Because
of the way it's structured, new metadata fields can easily be forgotten
and pose debugging challenges to developers. This likely warrants a new
issue.
I'm guessing that the implementation is done this way to avoid issues
with deserializing elements, but could be wrong.
---------
Co-authored-by: Benjamin Torres <benjats07@users.noreply.github.com>
2023-09-14 11:23:16 -04:00
|
|
|
--metadata-exclude filename,file_directory,metadata.data_source.date_processed,metadata.date,metadata.detection_class_prob,metadata.parent_id,metadata.category_depth \
|
2023-09-26 17:06:53 -04:00
|
|
|
--num-processes "$max_processes" \
|
2023-08-11 22:02:51 +03:00
|
|
|
--preserve-downloads \
|
|
|
|
|
--reprocess \
|
2023-09-11 11:40:56 -04:00
|
|
|
--output-dir "$OUTPUT_DIR"
|
2023-08-11 22:02:51 +03:00
|
|
|
|
|
|
|
|
|
|
|
|
|
# We are expecting fifteen directories: fourteen bases and the parent directory
|
2023-09-21 14:51:08 -04:00
|
|
|
"$SCRIPT_DIR"/check-num-dirs-output.sh 15 "$OUTPUT_FOLDER_NAME"
|
2023-08-11 22:02:51 +03:00
|
|
|
|
|
|
|
|
# We are expecting 101 files: 100 tables and the parent directory
|
2023-09-21 14:51:08 -04:00
|
|
|
"$SCRIPT_DIR"/check-num-files-output.sh 101 "$OUTPUT_FOLDER_NAME"/"$LARGE_BASE_BASE_ID"/
|
2023-08-11 22:02:51 +03:00
|
|
|
|
|
|
|
|
# Test on ingesting a large number of bases
|
|
|
|
|
for i in {1..12}; do
|
|
|
|
|
var="LARGE_WORKSPACE_BASE_ID_$i"
|
2023-09-21 14:51:08 -04:00
|
|
|
"$SCRIPT_DIR"/check-num-files-output.sh 12 "$OUTPUT_FOLDER_NAME"/"${!var}"
|
2023-08-11 22:02:51 +03:00
|
|
|
done
|
|
|
|
|
|
|
|
|
|
# Test on ingesting a table with lots of rows
|
2023-09-21 14:51:08 -04:00
|
|
|
"$SCRIPT_DIR"/check-num-rows-and-columns-output.sh 39999 "$OUTPUT_DIR"/"$LARGE_TABLE_BASE_ID"/"$LARGE_TABLE_TABLE_ID".json
|