Get element contents - Unstructured

Task

You want to get, manipulate, and print or save, the contents of the document elements and metadata from the processed data that Unstructured returns.

Approach

Each element in the document elements contains fields for that element’s type, its ID, the extracted text, and associated metadata.

The programmatic approach you take to get these document elements will depend on which tool, SDK, or library you use:

Ingest CLI

For the Unstructured Ingest CLI, you can use a tool such as jq to work with a JSON file that the CLI outputs after the processing is complete.

For example, the following script uses jq to access and print each element’s ID, text, and originating file name:

Shell

#!/usr/bin/env bash

JSON_FILE="local-ingest-output/my-file.json"

jq -r '.[] | "ID: \(.element_id)\nText: \(.text)\nFilename: \(.metadata.filename)\n"' \
    "$JSON_FILE"

Ingest Python library

For the Unstructured Ingest Python library, you can use the standard Python json.load function to load into a Python dictionary the contents of a JSON file that the Ingest Python library outputs after the processing is complete.

For example, the following code example uses standard Python to access and print each element’s ID, text, and originating file name:

Python

import json

def parse_json_file(input_file_path: str):
    with open(input_file_path, 'r') as file:
        file_elements = json.load(file)

    for element in file_elements:
        print(f"ID: {element["element_id"]}")
        print(f"Text: {element["text"]}")
        print(f"Filename: {element["metadata"]["filename"]}\n")

if __name__ == "__main__":
    parse_json_file(
        input_file_path="local-ingest-output/my-file.json"
    )

Open-source library

For the Unstructured open-source library, calling the partition_via_api function returns a list of elements (list[Element]). For example:

Python

# ...

elements = partition_via_api(
    filename=input_filepath,
    api_key=os.getenv("UNSTRUCTURED_API_KEY"),
    api_url=os.getenv("UNSTRUCTURED_API_URL"),
    strategy="hi_res"
)

# ...

You can use standard Python list operations on this list.

You can also use standard Python looping techniques on this list to access each element in this list.

Each individual element has the following attributes:

.text provides the element’s text field value as a str. See Element example.
.metadata provides the element’s metadata field as an ElementMetadata object. See Metadata.
.category provides the element’s type field value as a str. See Element type.
.id provides the element’s element_id value as a str. See Element ID.

In addition, the following methods are available:

.convert_coordinates_to_new_system() converts the element’s location coordinates, if any, to a new coordinate system. See Element’s coordinates.
.to_dict() gets the element’s content as a standard Python key-value dictionary (dict[str, Any]).

For example:

Python

# ...

elements = partition_via_api(
    filename=input_filepath,
    api_key=os.getenv("UNSTRUCTURED_API_KEY"),
    api_url=os.getenv("UNSTRUCTURED_API_URL"),
    strategy="hi_res"
)

for element in elements:
    # Do something with each element, for example:
    save_element_to_database(f"{element.id}")
    save_element_to_database(f"{element.text}")
    save_element_to_database(f"{element.metadata.filename}")

To serialize this list as a Python dictionary, you can use the elements_to_dicts method, for example:

Python

from unstructured.staging.base import elements_to_dicts

# ...

elements = partition_via_api(
    filename=input_filepath,
    api_key=os.getenv("UNSTRUCTURED_API_KEY"),
    api_url=os.getenv("UNSTRUCTURED_API_URL"),
    strategy="hi_res"
)

elements_dicts = elements_to_dicts(elements)

To serialize this list as JSON, you can use the elements_to_json function to convert the list of elements (Iterable[Element]) into a JSON-formatted string and then print or save that string. For example:

Python

from unstructured.staging.base import elements_to_json

# ...

elements = partition_via_api(
    filename=input_filepath,
    api_key=os.getenv("UNSTRUCTURED_API_KEY"),
    api_url=os.getenv("UNSTRUCTURED_API_URL"),
    strategy="hi_res"
)

json_elements = elements_to_json(
    elements=elements,
    indent=2
)

elements_to_json(
    elements=elements,
    indent=2,
    filename=output_filepath
)

On this page

Task
Approach

Task

You want to get, manipulate, and print or save, the contents of the document elements and metadata from the processed data that Unstructured returns.

Approach

Each element in the document elements contains fields for that element’s type, its ID, the extracted text, and associated metadata.

The programmatic approach you take to get these document elements will depend on which tool, SDK, or library you use:

Ingest CLI

For the Unstructured Ingest CLI, you can use a tool such as jq to work with a JSON file that the CLI outputs after the processing is complete.

For example, the following script uses jq to access and print each element’s ID, text, and originating file name:

Shell

#!/usr/bin/env bash

JSON_FILE="local-ingest-output/my-file.json"

jq -r '.[] | "ID: \(.element_id)\nText: \(.text)\nFilename: \(.metadata.filename)\n"' \
    "$JSON_FILE"

Ingest Python library

For example, the following code example uses standard Python to access and print each element’s ID, text, and originating file name:

Python

import json

def parse_json_file(input_file_path: str):
    with open(input_file_path, 'r') as file:
        file_elements = json.load(file)

    for element in file_elements:
        print(f"ID: {element["element_id"]}")
        print(f"Text: {element["text"]}")
        print(f"Filename: {element["metadata"]["filename"]}\n")

if __name__ == "__main__":
    parse_json_file(
        input_file_path="local-ingest-output/my-file.json"
    )

Open-source library

For the Unstructured open-source library, calling the partition_via_api function returns a list of elements (list[Element]). For example:

Python

# ...

elements = partition_via_api(
    filename=input_filepath,
    api_key=os.getenv("UNSTRUCTURED_API_KEY"),
    api_url=os.getenv("UNSTRUCTURED_API_URL"),
    strategy="hi_res"
)

# ...

You can use standard Python list operations on this list.

You can also use standard Python looping techniques on this list to access each element in this list.

Each individual element has the following attributes:

.text provides the element’s text field value as a str. See Element example.
.metadata provides the element’s metadata field as an ElementMetadata object. See Metadata.
.category provides the element’s type field value as a str. See Element type.
.id provides the element’s element_id value as a str. See Element ID.

In addition, the following methods are available:

.convert_coordinates_to_new_system() converts the element’s location coordinates, if any, to a new coordinate system. See Element’s coordinates.
.to_dict() gets the element’s content as a standard Python key-value dictionary (dict[str, Any]).

For example:

Python

# ...

elements = partition_via_api(
    filename=input_filepath,
    api_key=os.getenv("UNSTRUCTURED_API_KEY"),
    api_url=os.getenv("UNSTRUCTURED_API_URL"),
    strategy="hi_res"
)

for element in elements:
    # Do something with each element, for example:
    save_element_to_database(f"{element.id}")
    save_element_to_database(f"{element.text}")
    save_element_to_database(f"{element.metadata.filename}")

To serialize this list as a Python dictionary, you can use the elements_to_dicts method, for example:

Python

from unstructured.staging.base import elements_to_dicts

# ...

elements = partition_via_api(
    filename=input_filepath,
    api_key=os.getenv("UNSTRUCTURED_API_KEY"),
    api_url=os.getenv("UNSTRUCTURED_API_URL"),
    strategy="hi_res"
)

elements_dicts = elements_to_dicts(elements)

Python

from unstructured.staging.base import elements_to_json

# ...

elements = partition_via_api(
    filename=input_filepath,
    api_key=os.getenv("UNSTRUCTURED_API_KEY"),
    api_url=os.getenv("UNSTRUCTURED_API_URL"),
    strategy="hi_res"
)

json_elements = elements_to_json(
    elements=elements,
    indent=2
)

elements_to_json(
    elements=elements,
    indent=2,
    filename=output_filepath
)

On this page

Task
Approach

​Task

​Approach

FAQ

​Task

​Approach

Task

Approach

Task

Approach