docs: add examples, citation and other fixes #34
Conversation
alexandreteles
added
the
documentation
📝 WalkthroughWalkthroughThe changes in this pull request involve updates to the Changes
Possibly related PRs
Thank you for using CodeRabbit. We offer it for free to the OSS community and would appreciate your support in helping us grow. If you find it useful, would you consider giving us a shout-out on your favorite social media? 🪧 TipsChatThere are 3 ways to chat with CodeRabbit:
Note: Be mindful of the bot's finite context window. It's strongly recommended to break down tasks such as reading entire modules into smaller chunks. For a focused discussion, use review comments to chat about specific files and their changes, instead of using the PR comments. CodeRabbit Commands (Invoked using PR comments)
Other keywords and placeholders
CodeRabbit Configuration File (
|
alexandreteles
force-pushed
the
docs/examples
branch
from
21e1a08
to
05a9a05
Compare
alexandreteles
force-pushed
the
docs/examples
branch
from
a11c54f
to
8568c41
Compare
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Caution
Inline review comments failed to post. This is likely due to GitHub's limits when posting large numbers of comments.
Actionable comments posted: 7
🧹 Outside diff range and nitpick comments (6)
TikTok/Queries/Common.py (1)
108-113: Approval of enhanced error handling with suggestion for improvement.The modifications to the error handling for httpx.HTTPError demonstrate a marked improvement in error reporting capabilities. The inclusion of the response text in the error message provides crucial context for debugging API-related issues. The utilization of f-strings for error message construction is an efficient choice. The error is appropriately logged prior to raising the exception, and the original exception is correctly chained using the 'from' keyword, preserving the traceback.
While the changes are fundamentally sound, a minor enhancement could further improve the robustness of this error handling mechanism.
Consider implementing a safeguard against potential AttributeError in case the 'response' object is not available. Apply this diff to enhance the error handling:
except httpx.HTTPError as http_error: error_message = ( - f"HTTP error during API query: {http_error}.\nResponse: {response.text}" + f"HTTP error during API query: {http_error}.\n" + f"Response: {getattr(http_error, 'response', {}).text if hasattr(http_error, 'response') else 'No response available'}" ) logger.error(error_message) raise QueryException(error_message) from http_errorThis modification ensures that the error handling remains functional even if the 'response' object is not accessible, thereby increasing the robustness of the error reporting mechanism.
CHANGELOG.md (1)
11-11: Entry for refactor in v1.1.2 is adequately documented, with room for minor improvement.The changelog entry for the refactor in version 1.1.2 maintains consistency with the established format. It concisely states the change as modifying the exception handler. The commit hash (e1db066) is correctly included, facilitating traceability.
While the current description is sufficient, it could be enhanced by briefly mentioning the specific nature or purpose of the exception handler change, if such information would provide valuable context to readers of the changelog.
TikTok/ValidationModels/Video.py (4)
Line range hint
94-106: Consider Renaming 'MID' to 'MEDIUM' inVideoLengthEnumThe
VideoLengthenum definesSHORT,MID,LONG, andEXTRA_LONGas video length categories. To maintain clarity and consistency, consider renamingMIDtoMEDIUM. This change aligns with the full-word naming convention used in the other enum members.Apply this diff to update the enum member:
class VideoLength(StrEnum): """ Enum representing the available video length categories. Attributes: SHORT (str): Short video length category. - MID (str): Medium video length category. + MEDIUM (str): Medium video length category. LONG (str): Long video length category. EXTRA_LONG (str): Extra long video length category. """ SHORT = "SHORT" - MID = "MID" + MEDIUM = "MEDIUM" LONG = "LONG" EXTRA_LONG = "EXTRA_LONG"
Line range hint
689-691: Standardize Spelling offavorites_countThe field
favourites_countinVideoDataModeluses British English spelling. To maintain consistency with the rest of the codebase, which uses American English (e.g.,colorinstead ofcolour), consider renaming this field tofavorites_count.Apply this diff to update the field name:
class VideoQueryFields(StrEnum): """ Represents the possible fields for a video object. """ # ... - favourites_count = "favorites_count" + favorites_count = "favorites_count" # ... class VideoDataModel(BaseModel): """ Model representing the fields of a TikTok video search query. """ # ... - favourites_count: int | None = Field( + favorites_count: int | None = Field( default=None, description="The number of favorites that a video receives" )
Line range hint
153-165: Eliminate Redundant Enum DefinitionsThe
VideoQueryFieldNameenum duplicates the fields defined inVideoQueryFields. Maintaining two enums with identical values can lead to confusion and inconsistency. Consider consolidating these enums to a single definition to improve code maintainability.Apply this diff to remove the redundant enum:
class VideoQueryFields(StrEnum): """ Represents the possible fields for a video object. """ # Fields definition remains unchanged. - class VideoQueryFieldName(StrEnum): - """ - Enum representing the available field names for TikTok video queries. - """ - # Duplicate fields definition.
Line range hint
366-378: Correct Inaccurate Class References in DocstringsThe docstrings within the
VideoQueryRequestBuilderclass refer toTikTokQueryRequestBuilderinstead of the correct class name. This inconsistency may cause confusion for developers and hamper code readability.Apply this diff to correct the class references:
class VideoQueryRequestBuilder: """ - A builder class for constructing TikTokQueryRequest objects. + A builder class for constructing VideoQueryRequestModel objects. - This class provides methods to set various parameters of the TikTokQueryRequest + This class provides methods to set various parameters of the VideoQueryRequestModel and to add query conditions using the QueryBuilder. """
🛑 Comments failed to post (7)
TikTok/__init__.py (1)
19-19: 💡 Codebase verification
Incomplete Refactoring Detected.
The old import path
from TikTok.Types.OAuth2 import RequestHeadersModel, TokenRequestBodyModelis still present in the fileTikTok/Auth.py. This inconsistency can lead to potential import errors and disrupt the module restructuring effort.
- File with outdated import path:
TikTok/Auth.py🔗 Analysis chain
Import path modification detected.
The alteration in the import statement from
TikTok.Types.OAuth2toTikTok.ValidationModels.OAuth2forRequestHeadersModelandTokenRequestBodyModelclasses indicates a restructuring of the module organization. This modification appears to be part of a broader effort to categorize OAuth2-related validation models under a more semantically appropriate namespace.To ensure consistency across the codebase, execute the following verification script:
🏁 Scripts executed
The following scripts were executed for the analysis:
Script:
#!/bin/bash # Description: Verify the updated import path usage across the codebase # Test 1: Search for any remaining instances of the old import path echo "Searching for old import path usage:" rg "from TikTok.Types.OAuth2 import RequestHeadersModel, TokenRequestBodyModel" # Test 2: Confirm the presence of the new import path in other files echo "Confirming new import path usage:" rg "from TikTok.ValidationModels.OAuth2 import RequestHeadersModel, TokenRequestBodyModel" # Test 3: Check for any partial imports that might need updating echo "Checking for partial imports:" rg "from TikTok.Types.OAuth2 import (RequestHeadersModel|TokenRequestBodyModel)" rg "from TikTok.Types import OAuth2" # Test 4: Verify the existence of the new module structure echo "Verifying new module structure:" fd -t d "ValidationModels" TikTok fd -t f "OAuth2.py" TikTok/ValidationModelsLength of output: 1357
README.md (1)
20-20: 💡 Codebase verification
⚠️ Potential issueInconsistency Detected in Import Paths
The old import path
from TikTok.Types.OAuth2 import RequestHeadersModel, TokenRequestBodyModelremains inTikTok/Auth.py. This indicates an incomplete refactoring and may lead to import errors for users referencing the outdated module path.
- TikTok/Auth.py
🔗 Analysis chain
Import path modification detected.
The import statement for
RequestHeadersModelandTokenRequestBodyModelhas been altered fromTikTok.Types.OAuth2toTikTok.ValidationModels.OAuth2. This modification signifies a structural reorganization within the library, relocating these models from theTypesnamespace to theValidationModelsnamespace.This alteration may necessitate adjustments in existing user code. To verify the consistency of this change across the codebase and documentation, execute the following script:
Recommendation: Update the example code in the README to reflect this change, ensuring consistency between the import statement and the code examples provided.
🏁 Scripts executed
The following scripts were executed for the analysis:
Script:
#!/bin/bash # Description: Verify the consistency of import path changes across the codebase and documentation. # Test 1: Search for any remaining instances of the old import path echo "Searching for old import path..." rg "from TikTok.Types.OAuth2 import RequestHeadersModel, TokenRequestBodyModel" # Test 2: Verify the new import path is used consistently echo "Verifying new import path usage..." rg "from TikTok.ValidationModels.OAuth2 import RequestHeadersModel, TokenRequestBodyModel" # Test 3: Check for any mentions of the old path in comments or docstrings echo "Checking for mentions of old path in comments or docstrings..." rg "TikTok.Types.OAuth2"Length of output: 1110
examples/GetAllUserVideos.ipynb (5)
227-230: 🛠️ Refactor suggestion
Avoid using print statements within loops monitored by tqdm
The
tqdm. This can result in cluttered output and degraded readability of the progress indicators.Apply this diff to remove the
tqdmfeatures:for username in tqdm(usernames, desc="Processing Usernames", unit="username"): - print(f"Processing: {username}") user_videos = await fetch_videos(username, date_list, max_retries) all_videos.extend(user_videos)If you need to display additional information, consider updating the
tqdmdescription dynamically:for username in tqdm(usernames, desc="Processing Usernames", unit="username"): tqdm.write(f"Processing: {username}") ...
261-264: 🛠️ Refactor suggestion
Potential performance issues with large datasets
Converting a list of video models to a DataFrame using a list comprehension may lead to high memory usage if the dataset is large. This could potentially cause performance degradation or memory errors.
Consider processing the data in chunks or using generators to reduce memory consumption:
df = pd.DataFrame.from_dict( - [video.model_dump(by_alias=True, exclude_none=True) for video in videos] + (video.model_dump(by_alias=True, exclude_none=True) for video in videos) )By using a generator expression instead of a list comprehension, you can reduce memory overhead when creating the DataFrame.
📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements." (video.model_dump(by_alias=True, exclude_none=True) for video in videos)\n", ")\n", "\n", "df.fillna(0, inplace=True) # Replace NaN with 0. Optional.\n"
242-243:
⚠️ Potential issueUndefined variable 'profiles'
The variable
profilesis not defined prior to its usage, which will result in aNameErrorupon execution. This prevents the code from running successfully.Define the
profilesvariable or replace it with the correct variable name:# Remember to replace the list of profiles with your own list of usernames. -videos = await fetch_videos_for_usernames(profiles, dates) +usernames = ['username1', 'username2', 'username3'] # Replace with your own list of usernames. +videos = await fetch_videos_for_usernames(usernames, dates)Ensure that the list of usernames is correctly assigned to the variable used in the function call.
Committable suggestion was skipped due to low confidence.
19-46:
⚠️ Potential issuePotential missing ValueError when start date is after end date
The function
generate_date_listspecifies in its docstring that it raises aValueErrorif the start date occurs after the end date. However, the current implementation does not include this check, which could lead to unexpected behavior if the start date is after the end date.Apply this diff to add the missing check:
def generate_date_list(start_date: str, end_date: str) -> list[str]: """ Generate a list of dates between two dates, inclusive. ... """ date_format = "%Y%m%d" start_date = datetime.strptime(start_date, date_format) end_date = datetime.strptime(end_date, date_format) + if start_date > end_date: + raise ValueError("Start date must not be after end date.") delta = end_date - start_date return [ (start_date + timedelta(days=i)).strftime(date_format) for i in range(delta.days + 1) ]📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.def generate_date_list(start_date: str, end_date: str) -> list[str]: """ Generate a list of dates between two dates, inclusive. Args: start_date (str): The start date as a string in "YYYYMMDD" format. end_date (str): The end date as a string in "YYYYMMDD" format. Returns: list of str: A list containing all dates from the start date to the end date, inclusive, each formatted as "YYYYMMDD". Raises: ValueError: If the input date strings are not in the correct format or if the start date occurs after the end date. Example: >>> generate_date_list("20240811", "20240815") ['20240811', '20240812', '20240813', '20240814', '20240815'] """ date_format = "%Y%m%d" start_date = datetime.strptime(start_date, date_format) end_date = datetime.strptime(end_date, date_format) if start_date > end_date: raise ValueError("Start date must not be after end date.") delta = end_date - start_date return [ (start_date + timedelta(days=i)).strftime(date_format) for i in range(delta.days + 1) ]
161-166: 🛠️ Refactor suggestion
Catching broad exceptions may mask unexpected errors
Using a broad
except Exception as e:block can obscure unexpected errors and make debugging difficult. It's advisable to catch specific exceptions that you expect might occur during the API call, such as network-related errors or API-specific exceptions.Consider catching specific exceptions:
try: video_query_response = await query.video.search( request=current_request, fields=[ VideoQueryFields.id, ... ], ) -except Exception as e: +except NetworkError as e: + # Handle network-related exceptions + retries += 1 + if retries > max_retries: + break + await asyncio.sleep(2**retries) # Exponential backoff + continue +except ApiError as e: + # Handle API-specific exceptions + # Decide whether to retry or fail based on the error + breakReplace
NetworkErrorandApiErrorwith the specific exceptions that thequery.video.searchmethod might raise.Committable suggestion was skipped due to low confidence.
There was a problem hiding this comment.
Actionable comments posted: 1
🧹 Outside diff range and nitpick comments (7)
pyproject.toml (1)
27-31: Development dependencies added.New development dependencies
ipython(^8.28.0) andipykernel(^6.29.5) have been introduced. These additions align with the pull request objective of incorporating dependencies necessary for documentation.Consider updating the project's documentation to reflect the addition of these development dependencies and provide instructions for their usage in the development environment.
examples/GetAllUserVideos.ipynb (6)
1-9: Markdown cell provides essential API limitation information.The cell accurately describes the TikTok Research API's constraint of fetching a maximum of 100 videos per day. It appropriately recommends creating a list of dates for making requests, which is a logical workaround for this limitation. The inclusion of a reference to a GitHub issue for further information is commendable.
Consider adding a brief code snippet demonstrating the recommended date list creation to enhance user comprehension.
10-53: Functiongenerate_date_listis well-implemented with comprehensive error handling.The function effectively generates a list of dates between two given dates, utilizing the datetime module for precise date manipulation. The inclusion of error handling for incorrect date formats and invalid date ranges is commendable. The docstring provides clear documentation of the function's purpose, parameters, return value, and potential exceptions.
Consider adding input validation to ensure that the input strings match the expected "YYYYMMDD" format before attempting to parse them. This could be achieved using a regular expression check.
62-83: API authentication is correctly implemented using OAuth2.The code cell effectively demonstrates the authentication process using the OAuth2 class from the TikTok library. The use of environment variables for storing sensitive credentials is a commendable security practice. The initialization of the Query object for making subsequent requests is appropriate.
Consider adding error handling for cases where the environment variables are not set. This could prevent cryptic errors and provide more user-friendly feedback.
91-195: Functionfetch_videosdemonstrates robust implementation with error handling and pagination support.The function effectively fetches videos for a given username over multiple dates. Notable features include:
- Implementation of retry logic with exponential backoff, enhancing resilience against transient failures.
- Utilization of a progress bar for visual feedback, improving user experience.
- Proper handling of pagination using cursor and search_id parameters, ensuring comprehensive data retrieval.
- Comprehensive error handling and graceful degradation in case of persistent failures.
The function's structure and implementation align with best practices for interacting with paginated APIs.
Consider the following enhancements:
- Implement logging to capture details of retries and failures for debugging purposes.
- Add a parameter to allow users to customize the fields retrieved, providing more flexibility.
204-243: Functionfetch_videos_for_usernameseffectively implements multi-user video fetching.The function successfully extends the single-user functionality to handle multiple users. Key aspects include:
- Utilization of nested progress bars, providing clear visual feedback on the overall progress and individual user progress.
- Efficient leveraging of the previously defined
fetch_videosfunction, promoting code reuse.- Clear documentation of function parameters and return value.
The implementation is well-structured and aligns with the introduced concept of multi-user video fetching.
Consider implementing parallel processing using asyncio.gather() to fetch videos for multiple users concurrently, potentially improving overall execution time.
244-265: Data processing cell effectively demonstrates conversion to pandas DataFrame.The code cell successfully illustrates the process of converting the fetched video data into a pandas DataFrame. Key observations:
- Utilization of the
model_dumpmethod with appropriate parameters to convert video objects to dictionaries.- Creation of a DataFrame from the list of dictionaries.
- Optional step to replace NaN values with 0, which can be useful for certain types of analysis.
This example provides users with a clear path for further data manipulation and analysis using pandas.
Consider adding a brief comment explaining the potential implications of replacing NaN values with 0, as this may not be appropriate for all use cases.
📜 Review details
Configuration used: CodeRabbit UI
Review profile: CHILL
⛔ Files ignored due to path filters (1)
poetry.lockis excluded by!**/*.lock
📒 Files selected for processing (4)
- README.md (1 hunks)
- TikTok/init.py (1 hunks)
- examples/GetAllUserVideos.ipynb (1 hunks)
- pyproject.toml (1 hunks)
🚧 Files skipped from review as they are similar to previous changes (2)
- README.md
- TikTok/init.py
🧰 Additional context used
🔇 Additional comments (4)
pyproject.toml (1)
Line range hint
3-3: Version increment detected.The project version has been incremented from 1.1.2 to 1.1.6, adhering to semantic versioning principles. This change suggests the introduction of new features or non-breaking changes.
examples/GetAllUserVideos.ipynb (3)
55-61: Markdown cell promotes secure credential management.The cell appropriately recommends storing the client key and secret in environment variables. This practice aligns with security best practices by preventing the exposure of sensitive credentials in the code.
84-90: Markdown cell provides useful information on retry logic.The cell accurately informs users about the library's use of tenacity for handling retries. It appropriately suggests that users can implement their own retry logic if needed, which is valuable information for those requiring custom retry behavior.
196-203: Markdown cell effectively introduces multi-user video fetching concept.The cell succinctly introduces the concept of fetching videos from multiple users, providing appropriate context for the subsequent code implementation. This information is valuable for users who need to process data from multiple TikTok accounts.
pyproject.toml
Outdated
| @@ -16,14 +16,19 @@ rich = "13.9.2" | |||
| stamina = "24.3.0" | |||
| structlog = "24.4.0" | |||
| setuptools = "75.2.0" | |||
| cytoolz = "^1.0.0" | |||
| cytoolz = "^0.12.3" | |||
There was a problem hiding this comment.
💡 Codebase verification
Compatibility Concern with cytoolz Downgrade
The valfilter function from cytoolz is utilized in TikTok/Queries/Common.py. Downgrading cytoolz to version ^0.12.3 may lead to compatibility issues, as valfilter might be unsupported or altered in this version.
🔗 Analysis chain
Dependency version downgrade detected.
The cytoolz dependency has been downgraded from version ^1.0.0 to ^0.12.3. This major version downgrade may introduce compatibility issues or remove features utilized in the current codebase.
Execute the following script to identify potential compatibility issues:
🏁 Scripts executed
The following scripts were executed for the analysis:
Script:
#!/bin/bash
# Description: Search for cytoolz usage in the codebase to identify potential compatibility issues.
# Test: Search for cytoolz imports and usage
rg -i 'from cytoolz|import cytoolz|cytoolz\.' -g '!pyproject.toml'
Length of output: 123
There was a problem hiding this comment.
Actionable comments posted: 1
🧹 Outside diff range and nitpick comments (1)
README.md (1)
91-104: Citation section addition: Approved with minor formatting suggestion.The incorporation of a "Citation" section (lines 91-104) aligns with established academic software documentation practices. The provided BibTeX entry encompasses all requisite information for accurate citation, including author, title, year, DOI, URL, and institution. The strategic placement at the document's conclusion is logically sound.
To enhance the visual demarcation of the BibTeX entry, consider enclosing it within a fenced code block with a language specifier. Apply the following modification:
# Citation If you use this library in your work, please cite it as follows: -``` +```bibtex @misc{teles_tiktok_sdk_2024, author = {Alexandre Teles}, title = {{TikTok Research API SDK}}, year = 2024, doi = {10.5281/zenodo.13968316}, url = {https://github.com/INCT-DD/tiktok-sdk}, institution = {Instituto Nacional de Ciência e Tecnologia em Democracia Digital}, }<details> <summary>🧰 Tools</summary> <details> <summary>🪛 Markdownlint</summary><blockquote> 95-95: null Fenced code blocks should have a language specified (MD040, fenced-code-language) </blockquote></details> </details> </blockquote></details> </blockquote></details> <details> <summary>📜 Review details</summary> **Configuration used: CodeRabbit UI** **Review profile: CHILL** <details> <summary>📥 Commits</summary> Files that changed from the base of the PR and between 8568c41dcd568fcddaa3b79030f58d4c1d43cf00 and 4eeccf0e9cfaf13409b0d33c50ae1b6890fcd26a. </details> <details> <summary>📒 Files selected for processing (2)</summary> * README.md (3 hunks) * TikTok/__init__.py (3 hunks) </details> <details> <summary>🚧 Files skipped from review as they are similar to previous changes (1)</summary> * TikTok/__init__.py </details> <details> <summary>🧰 Additional context used</summary> <details> <summary>🪛 Markdownlint</summary><blockquote> <details> <summary>README.md</summary><blockquote> 95-95: null Fenced code blocks should have a language specified (MD040, fenced-code-language) </blockquote></details> </blockquote></details> </details> <details> <summary>🔇 Additional comments (2)</summary><blockquote> <details> <summary>README.md (2)</summary><blockquote> `3-4`: **DOI badge integration: Approved.** The insertion of the DOI badge at line 3 adheres to established practices for academic software documentation. Its prominent placement at the document's inception ensures optimal visibility, facilitating proper citation and impact tracking of the software artifact. --- Line range hint `1-104`: **README structure and content: Comprehensive and well-organized.** The README document exhibits a logically structured presentation of the TikTok library. It encompasses: 1. An introductory overview elucidating the library's purpose and technological foundation. 2. Detailed usage examples demonstrating authentication and query processes. 3. Academic citation information, including a DOI badge and BibTeX entry. This composition provides potential users with a clear understanding of the library's functionality and proper attribution methods. <details> <summary>🧰 Tools</summary> <details> <summary>🪛 Markdownlint</summary><blockquote> 95-95: null Fenced code blocks should have a language specified (MD040, fenced-code-language) </blockquote></details> </details> </blockquote></details> </blockquote></details> </details> <!-- This is an auto-generated comment by CodeRabbit for review status -->
| @@ -17,7 +19,7 @@ of the Auth class with your API key and secret: | |||
|
|
|||
| ```python | |||
| from TikTok.Auth import OAuth2 | |||
| from TikTok.Types.OAuth2 import RequestHeadersModel, TokenRequestBodyModel | |||
| from TikTok.ValidationModels.OAuth2 import RequestHeadersModel, TokenRequestBodyModel | |||
There was a problem hiding this comment.
💡 Codebase verification
Incomplete Refactoring: Old Import Path Detected in TikTok/Auth.py
The import statement from TikTok.Types.OAuth2 import RequestHeadersModel, TokenRequestBodyModel remains in TikTok/Auth.py, indicating that the refactoring to TikTok.ValidationModels.OAuth2 is incomplete.
- File requiring update:
TikTok/Auth.py
🔗 Analysis chain
Import path modification: Approved with verification recommendation.
The alteration of the import path from TikTok.Types.OAuth2 to TikTok.ValidationModels.OAuth2 for RequestHeadersModel and TokenRequestBodyModel is logically sound and maintains consistency with observed changes in TikTok/__init__.py. This modification reflects a deliberate restructuring of the module organization.
To ensure comprehensive implementation of this structural change, execute the following verification script:
🏁 Scripts executed
The following scripts were executed for the analysis:
Script:
#!/bin/bash
# Description: Verify the updated import path usage across the codebase.
# Test: Search for any remaining instances of the old import path.
rg -i "from TikTok.Types.OAuth2 import RequestHeadersModel, TokenRequestBodyModel"
# Test: Confirm the presence of the new import path in relevant files.
rg -i "from TikTok.ValidationModels.OAuth2 import RequestHeadersModel, TokenRequestBodyModel"
Length of output: 708
alexandreteles
changed the title
Summary by CodeRabbit
New Features
Bug Fixes
Documentation