> ## Documentation Index
> Fetch the complete documentation index at: https://docs.bigdata.com/llms.txt
> Use this file to discover all available pages before exploring further.
# bigdata-client Changelog
We are sunsetting our SDKs and will no longer add new features, security patches, bug fixes, or technical support for them. To access the latest capabilities and ongoing improvements, we encourage you to migrate to our [RESTful API](/api-rest/introduction).
SDK support will officially end on **December 31, 2026**. On this date, the underlying endpoints used by the SDKs and related documentation will be decommissioned.
To avoid any disruption to your services, please ensure your migration is complete by that date.
For migration assistance, please contact us at [support@bigdata.com](mailto:support@bigdata.com).
All notable changes to the [bigdata-client](https://pypi.org/project/bigdata-client/) package will be documented in this file.
The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/), and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
## \[2.20.0] - 2026-04-10
**Added**
* `FilingTypes.SEC_DEF_14A` for definitive proxy statements (SEC Form DEF 14A), aligned with the Search API [`document_type` filing subtypes](../getting-started/search/query_filters#filing).
## \[2.19.1] - 2026-01-15
**Fixed**
* Paginating uploaded files raised errors with large numbers.
## \[2.19.0] - 2025-09-24
**Changed**
* Search usage values are now floats instead of integers, so customers can see the usage with more precision. Check [Monitor usage](../how-to-guides/monitor_usage#per-request) for more details.
**Fixed**
* Print an error message to inform that the service `bigdata.uploads` does not support API KEY authentication yet.
## \[2.18.2] - 2025-09-01
**Added**
* Subscribers can view the usage of each Chat interaction by calling `usage` from the response object.
```python theme={null}
response = chat.ask("Speakers at the Figma earnings call?")
# Chat interaction usage
print(f"Input tokens: {response.usage.input_tokens}")
print(f"Output tokens: {response.usage.output_tokens}")
print(f"Searches count: {response.usage.searches_count}")
```
Output:
```console theme={null}
Input tokens: 19
Output tokens: 953
Searches count: 3
```
## \[2.18.1] - 2025-08-11
**Fixed**
* Chat Service response sources now contain unique entries, so Inline attributions do not have duplicate references to the same response source ID. Check out the how-to guide [Inline Attribution Details](../how-to-guides/chat/inline_attribution_details) for more details.
## \[2.18.0] - 2025-07-18
**Added**
* Enhance the Knowledge Graph's `find_sources` method to return the favicon URL.
* Chat response sources details: timestamp, and thumbnail images, so clients can display inline attribution sources like in Bigdata App:
Check out the how to guide [Inline Attribution Details](../how-to-guides/chat/inline_attribution_details) to learn how to get all the inline attribution details, a favicon and download the entire document.
For limited customers only:
* Beta support of [API Keys](../sdk-reference/introduction#api-key)
* Beta support of Remote MCP server to get transcripts insights using the Search Service. Check [MCP Reference](../mcp-reference/introduction#transcripts-search-mcp) for more details.
**Fixed**
* Implemented re-try logic for the `chat.ask` method to prevent the following error when establishing the connection:
```console theme={null}
Failed: Chat timeout (10.01 seconds) for scope None: timed out while waiting for handshake response, number of msg we have received are: 0, last msg is: No messages received
```
**Removed**
* The Watchlist query filter is no longer supported. However, customers can still create Watchlists and craft a query with the list of entities and the operator `Any`. Please see an example in [Query filters > Watchlist](../getting-started/search/query_filters#watchlist).
## \[2.17.0] - 2025-07-03
**Added**
* The method `chat.ask` now supports source filtering. Check the parameter `source_filter` in the [Chat API Reference](../sdk-reference/chat/chat#param-source-filter) for more details.
## \[2.16.1] - 2025-06-12
**Added**
* [find\_companies](../getting-started/knowledge_graph/find_companies#find-companies-by-details) can also be used with an empty name value to find the most relevant companies from a specific country.
## \[2.16.0] - 2025-06-11
**Added**
* Enhanced [find\_companies](../getting-started/knowledge_graph/find_companies#find-companies-by-details) with type and country filters.
## \[2.15.0] - 2025-06-05
**Added**
* Enhanced filtering for finding sources. Check [Find Sources](../getting-started/knowledge_graph/find_sources) for more details.
* Introduced a new method to wait for the file to be fully processed `file.wait_for_analysis_complete()`. Check [Upload Content](../getting-started/upload_content) for more details.
* ChatService now supports two different [ChatScope](../sdk-reference/chat/chatScope) to narrow responses to transcripts sources.
* `ChatScope.EARNINGS_CALL`: It uses transcripts from the Quartr subscription.
* `ChatScope.FACTSET_TRANSCRIPTS`: It uses transcripts from the FactSet subscription.
**Fixed**
* The method `file.wait_for_completion()` now waits for the file to be fully processed, including the indexing step if required. Check [Upload Content](../getting-started/upload_content) for more details.
* The default wait time is `2400`seconds (40 minutes), but you can customize it with the `timeout` parameter (In seconds).
* The date range in [AbsoluteDateRange](../getting-started/search/date_range) is more robust.
## \[2.14.0] - 2025-05-21
**Added**
* Proxy support for WebSocket connections. Check [Proxy configuration](../sdk-reference/proxy_configuration) for more details.
**Fixed**
* Fixed a bug in async apps where the async lib was being patched at a global scope.
* Fixed a bug reading Chat responses
```python theme={null}
Field required [type=missing, input_value={'type': 'COMPLETE', 'req..., 'contentBlocks': None}, input_type=dict]
```
## \[2.13.0] - 2025-04-23
**Changed**
* Request to Transcript and Filing types generate new request parameter
**Added**
* New search class `DocumentVersion`
* Tracking parameter for `chat.ask` requests
## \[2.12.0] - 2025-04-09
**Added**
* The method `chat.ask` supports streaming with the parameter `streaming`. More details in [Chat API Reference](../sdk-reference/chat/chatService)
* Filter watchlist by `owned` parameter
## \[2.11.1] - 2025-03-28
**Changed**
* HTTP header support
## \[2.11.0] - 2025-03-27
**Added**
* Inline attributions to chat responses
* Inline attribution formatters: DefaultInlineAttributionFormatter
* Support for sort by date in `asc` order. `SortBy.DATE_ASC`
## \[2.10.1] - 2025-03-17
**Fixed**
* Typo in ChatScope
## \[2.10.0] - 2025-03-13
**Added**
* Added `scope` parameter to `Chat.ask` method
* Method to find ETFs in the [Knowledge Graph](../getting-started/knowledge_graph/find_etfs): `find_etfs`
## \[2.9.1] - 2025-03-04
**Fixed**
* Creating a new Clerk session from scratch when an error happens during the refreshing process.
## \[2.9.0] - 2025-02-26
**Added**
* Chat experience via API
* Quickstart guide: [Chat Service](../getting-started/quickstart_guide_chat)
* [Chat Service API](../sdk-reference/chat/chatService)
* Enhance Bigdata class to manage rate limits, optimizing bandwidth.
* Enhance autosuggest response with company market Identifiers
* Exception `BigdataClientIncompatibleStateError` is raised when trying to tag or share a file which upload and classification process is not COMPLETED
* New exception `BigdataClientSimilarityPayloadTooLarge` to notify users when the input from a Similary query is too large.
## \[2.8.0] - 2025-02-13
**Added**
* Feature to query by chunks instead of by documents. Specify more precise limits to your queries. [ChunkLimit](../how-to-guides/queries_with_chunk_limit)
**Fixed**
* In multithreading scenarios, more than one thread could start a refreshing token request.
## \[2.7.1] - 2025-02-03
**Added**
* LICENCE file
## \[2.7.0] - 2025-01-30
**Added**
* Added `rerank_threshold` param to `bigdata_client.search.new`.
* Limit before the client raises `RequestMaxLimitExceeds` raised from 8KB -> 64KB.
* Support for downloading very large documents (> 6MB) with `bigdata_client.document.download_annotated_dict` method.
* If a method to delete a file is called and it has not been fully processed yet it will raise `BigdataClientIncompatibleStateError`.
## \[2.6.0] - 2025-01-17
**Added**
* Added methods for checking the tags of your files: `bigdata_client.uploads.list_my_tags`. ([List my tags](../getting-started/upload_content#list-my-tags).
* Added methods for checking the tags of files shared with you: `bigdata_client.uploads.list_tags_shared_with_me`. ([List tags shared with me](../getting-started/upload_content#list-tags-shared-with-me)).
* Added methods for company retrieval as part of the [Knowledge Graph service](../getting-started/knowledge_graph/introduction):
* `get_companies_by_isin`
* `get_companies_by_cusip`
* `get_companies_by_sedol`
* `get_companies_by_listing`
* The response from `subscription.get_details()` will include information about uploaded pages of PDF files: `pdf_upload_pages`. Check [Monitor usage](../how-to-guides/monitor_usage).
**Fixed**
* Retrieved watchlists were not setting the `company_shared_permission` field.
* The response from `subscription.get_details()` will correctly inform about uploaded pages of files (other than PDF): `file_upload_pages`.
* Scripts `batch_file_analytics_download.py` and `batch_file_upload.py` now correctly handles `BigdataClientRateLimitError`.
## \[2.5.0] - 2024-12-20
**Added**
* Added new query component `FileTag`, [usage here](../getting-started/search/query_filters#filetag).
* Added exceptions for controlling Authentication errors: `BigdataClientAuthFlowError`, `BigdataClientTooManySignInAttemptsError`
## \[2.4.0] - 2024-11-26
**Added**
* Enhance API [usage monitoring](../how-to-guides/monitor_usage) for private files uploads with `bigdata_client.subscription.get_details()` method.
* Added methods `add_tags`, `remove_tags` and `set_tags` for `File` class objects. These methods allow modifying the tags on existing files.
* Added pagination to `bigdata.upload.Uploads.list` method to retrieve all the files uploaded by the user.
Example of usage:
```python theme={null}
from bigdata_client import Bigdata
bigdata = Bigdata()
# New usage to retrieve all of your files
for n in itertools.count(start=1):
files = bigdata_cli.uploads.list(page_number=n)
do_stuff_with_files(files)
if not files:
break
```
**Fixed**
* Validation errors when parsing entities of type `bigdata_client.models.entities.Concept` in the knowledge graph service since v2.1.0.
* Entities of type `bigdata_client.models.entities.Topic` were incorrectly parsed as `bigdata_client.models.entities.Concept`.
* Validation errors when parsing edge cases of entities of type `bigdata_client.models.sources.Source` in the knowledge graph service.
## \[2.3.0] - 2024-11-15
**Added**
* Add new method `subscription.get_details()` for API usage monitoring.
* Add new method `get_usage()` at the class `bigdata_client.search.Search` that returns the API Query units used for each search instance.
* Enhance exception handling. Check [Exceptions](../sdk-reference/exceptions).
## \[2.2.0] - 2024-10-30
**Added**
* Enhance error handling experience. Customers will see error messages besides the HTTP code.
* Enhance methods in the knowledge\_graph service to provide more concise answers.
**Changed**
* The following methods from the knowledge\_graph service: `autosuggest`, `find_concepts`, `find_companies`, `find_people`,
`find_places`, `find_organizations`, `find_products`, `find_sources` and `find_topics` should now be used with a single
parameter instead of a list. E.g:
Example of current usage:
```python theme={null}
from bigdata_client import Bigdata
bigdata = Bigdata()
bigdata.knowledge_graph.autosuggest(["Company 1", "Company 2"])
```
After the update:
```python theme={null}
from bigdata_client import Bigdata
bigdata = Bigdata()
bigdata.knowledge_graph.autosuggest("Company 1")
bigdata.knowledge_graph.autosuggest("Company 2")
```
The old implementation is now flagged as `deprecated` and compatibility with it will be removed in the future. This update
fixes problems from clients using the methods above when using concurrency.
A more detailed guide is included in the [knowledge graph documentation](../getting-started/knowledge_graph/introduction) and [how to guides](../how-to-guides/threading_find_entities).
**Fixed**
Enhance methods in the knowledge\_graph service to avoid errors when customers explore the Knowledge Graph using a multithreading environment.
## \[2.1.0] - 2024-10-09
**Added**
* Added `url` property to `Document` class
* Queries using `AbsoluteDateRange` accept now to be created with a timezone.
* Added validation for `post`, `patch` and `put` requests that json request body not exceeds 8KB
* Added `verify_ssl` parameter for `bigdata_client.Bigdata` class to be able to skip ssl verification for proxy
## \[2.0.0] - 2024-09-19
**Changed**
* The name of the Python package changes from `bigdata` to `bigdata_client`.
We are doing this to ensure that there are no conflicts with other commonly used Python packages.
This change will require action on your part.
What you need to do:
Currently, you import classes from the package `bigdata`.
Once you update the package to version 2.0.0, be sure to modify your Python scripts to import
classes from `bigdata_client` instead of `bigdata` to avoid any issues.
Example of current imports:
```python theme={null}
from bigdata import Bigdata
from bigdata.query import Entity, TranscriptTypes
```
After the update:
```python theme={null}
from bigdata_client import Bigdata
from bigdata_client.query import Entity,TranscriptTypes
```
## \[1.6.1] - 2024-09-13
**Added**
* Added `download_annotated_dict` method to `Document` class
## \[1.6.0] - 2024-09-02
**Added**
* New environment to customize the number of parallel requests that can be made
to the API. The environment variable is `BIGDATA_MAX_PARALLEL_REQUESTS`.
* Added new parameter `skip_metadata: Optional[bool]` to `Uploads.upload_from_disk`, it allow skip loading file metadata
when uploading file.
## \[1.5.1] - 2024-08-26
**Changed**
* Expose FilingTypes in `bigdata.query`
## \[1.5.0] - 2024-08-26
**Added**
* New query component FilingTypes. Allow filter the current SEC Filings types.
## \[1.4.1] - 2024-08-19
**Changed**
* Internal components used to parse knowledge graph responses have been moved to the `/api` module.
## \[1.4.0] - 2024-08-12
**Added**
* New feature: Define your [primary entity](../getting-started/upload_content) when uploading a file.
* Knowledge graph `get_entities()`, `get_sources()`, and `get_topics()` now return analytic descriptions when available.
* The environment variable used to login into Bigdata `BIGDATA_USER` has been renamed to `BIGDATA_USERNAME`. Old version
is still supported but marked as deprecated.
**Changed**
* Improved keyword search capability to make phrase searches more accurate.
* Comention searches respect the date ranges defined in `Bigdata.search.new()`.
## \[1.3.0] - 2024-07-17
**Added**
* New key **cluster** can now be present in the `Document` object. It contains related `Document` objects.
* New method `search.find_concepts` to return the first concepts from the autosuggest service.
* Improved query import. Import `Company`, `Concept`, `Facility`, `Landmark`, `Organization`, `OrganizationType`, `Person`, `Place`, `Product`, `ProductType` from bigdata.
* Added timezone information to the datetimes in files.
**Fixed**
* There was an error when working with queries that have an empty list as the value for the field `date`.
## \[1.2.0] - 2024-06-26
**Added**
* New key **cluster** can now be present in the `Document` object. It contains related `Document` objects.
* Support for custom proxies.
## \[1.1.0] - 2024-06-10
**Added**
* Share/Unshare files directly from the `File` object or using `bigdata.uploads` new methods.
* File objects now have property `company_shared_permission` to know if they are being shared or not.
* Files shared with the user can now be retrieved with `bigdata.uploads.list_shared`.
## \[1.0.1] - 2024-06-07
**Fixed**
* Queries with Watchlists were not displaying results and taking a very long time to complete.
## \[1.0.0] - 2024-06-06
**Added**
* New feature, share/unshare watchlists.
* Feature: New expression for queries: ReportingEntity.
* Feature: New expression for queries: SentimentRange
**Changed**
* Watchlists now contain field `company_shared_permission` to know if they are being shared or not.
* Shared queries now return `SharePermission.READ` instead of `SearchSharePermission.READ`
* Attribute `Bigdata.content_search` renamed to `Bigdata.search`
* Method `ContentSearch.new_from_query` renamed to `ContentSearch.new`
* Helper functions to group query expressions `any_`, `all_` renamed to `Any`, `All`
## \[0.4.0] - 2024-05-28
**Added**
* Document query component available for making queries and running searches. e.g: `Document("BFA16B80ED117EAA5693E8BA")`
* Knowledge Graph filters: `find_companies, find_people, find_places, find_organizations, find_products, find_sources, find_topics`.
* Method for getting results from a search as a list -> `Search.run`
* New field *volume* is now returned for each entity in comentions.
* New method *to\_dict* for formatting a comentions object to a dictionary.
**Changed**
* Renamed `DocumentTypes` -> `TranscriptTypes`.
* Renamed `FileType` -> `DocumentType`.
* TranscriptTypes enum now can be used to create queries. e.g: `TranscriptTypes.EARNINGS_CALL`
* SectionMetadata enum can now be used to create queries. e.g: `SectionMetadata.QUESTION`
* Renamed `Search.limit_stories` -> `Search.limit_documents`
* Renamed `Story` -> `Document` as well as all of its components to be named like "Document"
* Improved string representation of a `Document`
**Fixed**
* Key `people` was not being returned by comentions.
## \[0.3.0] - 2024-05-10
**Added**
* Added `Search.share_with_company` and `Search.unshare_with_company` to share and
"unshare" a search with the user's company (\[#11]).
* Added method `File.get_analytics_dict` to get the analytics directly in
memory, as a dictionary (\[#14]). That is consistent with other methods like
`File.get_annotated_dict`.
* New expressions to interact with transcripts: SectionMetadata, DocumentType, FiscalQuarter, FiscalYear.
* Added chunk-level sentiment to objects in `Story.chunks` (\[#13]).
**Fixed**
* Parsing backend response from discovery-panel endpoint when there are no results raised an error.
**Changed**
* Changed the range of `Story.sentiment` and `Chunk.sentiment` to be between -1
and 1 (\[#13]).
## \[0.2.0] - 2024-05-03
**Added**
* The chunks in the search results now have a `relevance` attribute, which is a
float bigger than 0. You can access it on `story.chunks[x].relevance` (\[#10]).
**Removed**
* Removed the simplified version of the search
`Bigdata.content_search.new_search`, in favor of the more powerful
`Bigdata.content_search.new_from_query` (\[#9]).
**Fixed**
* Fixed a bug where the `scope` parameter of
`Bigdata.content_search.new_from_query` was being ignored, causing the search
to always have a scope `FileType.ALL` (\[#12]).
## \[0.1.0] - 2024-04-26
**Added**
* Implemented `Bigdata.internal_content` to support interacting with internal
content (uploading documents, listing, getting, deleting, downloading
original document/annotations/analytics, etc) (\[#7]).
## \[0.0.8] - 2024-04-22
**Added**
* Implemented `Bigdata.watchlists` to create, update, delete get and list
watchlists (\[#5]).
**Fixed**
* Autosuggest API response now makes use of a discriminator.
## \[0.0.7] - 2024-04-17
**Fixed**
* Internal change to stop showing emails in the description in pypi
* Error when receiving some Product objects in search responses, some of which
raised a `ValidationError`.
## \[0.0.6] - 2024-04-12
**Fixed**
* Fixed a bug using autosuggest in a jupyter notebook, caused by asyncio
(\[#4]).
## \[0.0.5] - 2024-04-11
**Added**
* New way to inspect the knowledge graph and get autosuggestions on
entities, topics, etc. (\[#2]).
* New way of searching and creating queries, using And/Or/Not operators
and the method `new_from_query` from the `Search` class. (\[#3]).
**Changed**
* Sentiment added to `Story.__str__` so it's printed when using `print(story)`
* Updated the content search to use the **breaking change** introduced in the
latest API version. This means that older versions of the package won't work
with the latest versions of the API
## \[0.0.4] - 2024-03-22
**Added**
* Added missing parameter `sortby` to the `new_search` method
* The search results (stories) now implement the `__str__` to have a better
looking output.
**Changed**
* Changed endpoint used for searching to have more accurate search results
* Changed the authentication used internally to Clerk (\[#1]).
* Method `get_related` renamed to `get_comentions`
* `AbsoluteDateRange`'s `__init__` now accepts both types `datetime` and `str`.
**Removed**
* The classmethod `from_strings` has been removed in favor of `__init__`
passing strings.
## \[0.0.3] - 2024-03-20
**Fixed**
* Fixed crash on Python 3.10 or lower, caused by a wrong import. Did not affect
Python 3.11 and above.
## \[0.0.2] - 2024-03-19
**Changed**
* Enum items are now uppercase
## \[0.0.1] - 2024-03-19
**Added**
* First public release.
* Basic search functionality via `bigdata.content_search.new_search()`.
* CRUD operations for saved searches
* Co-mentions
* Authentication via username and password