Yandex AI Studio SDK

https://github.com/yandex-cloud/yandex-ai-studio-sdk
Admin
A Python SDK that provides simple and efficient abstractions for interacting with Yandex Cloud AI...
Tokens:228,986
Snippets:1,840
Trust Score:7.9
License:Apache-2.0
Update:2 weeks ago
Show doc for...
Tokens:
Raw
### Install documentation dependencies

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/CONTRIBUTING.md

Install Sphinx and other required packages for building documentation. Run this in the repo root.

```bash
pip install -r docs/requirements.txt -e .
```

--------------------------------

### Installation with optional extras

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/src/yandex_ai_studio_sdk/cli/search_index/README.md

Install the base package and optional extras for different data sources. Features are split into optional extras to reduce dependencies.

```bash
# Base installation with local files and Confluence support
pip install yandex-ai-studio-sdk

# With Amazon S3 / Yandex Object Storage support
pip install "yandex-ai-studio-sdk[cli-s3]"

# With Wikipedia / MediaWiki support
pip install "yandex-ai-studio-sdk[cli-wiki]"

# All extras
pip install "yandex-ai-studio-sdk[cli-wiki,cli-s3]"
```

--------------------------------

### Install yandex-ai-studio-sdk with langchain extra

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/README.md

Install the SDK with the langchain extra to enable LangChain integration.

```sh
pip install yandex-ai-studio-sdk[langchain]
```

--------------------------------

### Install yandex-ai-studio-sdk with pip

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/README.md

Install the SDK via pip.

```sh
pip install yandex-ai-studio-sdk
```

--------------------------------

### Install CLI extras and run yandex-ai-studio vector-stores

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/README.md

Install the CLI with optional extras and run the vector-stores command to build an index from local PDF files.

```bash
# Install base CLI (local files + Confluence)
pip install yandex-ai-studio-sdk

# With S3 / Yandex Object Storage support
pip install "yandex-ai-studio-sdk[cli-s3]"

# With Wikipedia / MediaWiki support
pip install "yandex-ai-studio-sdk[cli-wiki]"

yandex-ai-studio vector-stores local docs/*.pdf --name "My index"
```

--------------------------------

### Python int method examples: bit_length, bit_count, as_integer_ratio

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/internals/bases.rst

Python interactive console examples showing the behavior of int methods bit_length(), bit_count(), and as_integer_ratio(). Each example demonstrates the method's output for given integer inputs.

```pycon
>>> bin(37)
'0b100101'
>>> (37).bit_length()
6
```

```pycon
>>> bin(13)
'0b1101'
>>> (13).bit_count()
3
```

```pycon
>>> (10).as_integer_ratio()
(10, 1)
>>> (-10).as_integer_ratio()
(-10, 1)
>>> (0).as_integer_ratio()
(0, 1)
```

--------------------------------

### Basic SDK usage with AIStudio

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/README.md

Basic usage example: create an AIStudio client, configure a completion model, run it, and iterate over the result.

```python
from yandex_ai_studio_sdk import AIStudio

sdk = AIStudio(folder_id="...", auth="<APIKey/IAMToken/SomethingElse>")

model = sdk.models.completions('yandexgpt')
model = model.configure(temperature=0.5)
result = model.run("foo")

for alternative in result:
    print(alternative)
```

--------------------------------

### wiki subcommand examples

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/src/yandex_ai_studio_sdk/cli/search_index/README.md

Examples of using the wiki subcommand: indexing a single page, multiple pages, exporting as Markdown, and accessing a private wiki with credentials. The --export-format option accepts text, html, or markdown (default text). Credentials can be provided via --username and --password or environment variables WIKI_USERNAME and WIKI_PASSWORD.

```bash
# Index a single Wikipedia page
yandex-ai-studio vector-stores wiki https://en.wikipedia.org/wiki/Machine_learning
```

```bash
# Index multiple pages
yandex-ai-studio vector-stores wiki \
  https://en.wikipedia.org/wiki/Machine_learning \
  https://en.wikipedia.org/wiki/Neural_network \
  https://en.wikipedia.org/wiki/Transformer_(deep_learning_architecture)
```

```bash
# Export content as Markdown
yandex-ai-studio vector-stores wiki \
  "https://en.wikipedia.org/wiki/Python_(programming_language)" \
  --export-format markdown
```

```bash
# Access a private wiki with credentials
yandex-ai-studio vector-stores wiki \
  https://wiki.example.com/wiki/Internal_docs \
  --username alice \
  --password secret
```

--------------------------------

### local subcommand examples

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/src/yandex_ai_studio_sdk/cli/search_index/README.md

Examples of the local subcommand: indexing a single file, multiple files, using shell globs, and specifying a custom index name with --name.

```bash
# Index a single file
yandex-ai-studio vector-stores local report.pdf

# Index multiple files
yandex-ai-studio vector-stores local docs/intro.txt docs/guide.md

# Use shell glob to include all .txt and .md files
yandex-ai-studio vector-stores local sample_docs/*.txt sample_docs/*.md

# Specify a custom name for the index
yandex-ai-studio vector-stores local report.pdf --name "Q4 Report"
```

--------------------------------

### as_integer_ratio() example

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/async/speechkit/domain.rst

Shows the output of the as_integer_ratio() method for integers 10, -10, and 0, returning the integer ratio pairs.

```pycon
>>> (10).as_integer_ratio()
(10, 1)
>>> (-10).as_integer_ratio()
(-10, 1)
>>> (0).as_integer_ratio()
(0, 1)
```

--------------------------------

### LangChain integration with AIStudio

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/README.md

LangChain integration example: create a LangChain-compatible model and invoke it with a list of messages.

```python
from yandex_ai_studio_sdk import AIStudio
from langchain_core.messages import AIMessage, HumanMessage

sdk = AIStudio(folder_id="...", auth="<APIKey/IAMToken/SomethingElse>")

model = sdk.models.completions('yandexgpt').langchain()

langchain_result = model.invoke([
    HumanMessage(content="hello!"),
    AIMessage(content="Hi there human!"),
    HumanMessage(content="Meow!"),
])
```

--------------------------------

### Configure STT audio format

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/async/speechkit/speech_to_text.rst

Shows how to change the initial speech-to-text settings using the `.configure` method. The example creates an STT instance with `audio_format='mp3'` and then reconfigures it to `audio_format='WAV'`.

```pycon
>>> stt = sdk.speechkit.speech_to_text(audio_format='mp3')
>>> stt = stt.configure(audio_format='WAV')
```

--------------------------------

### Configure generative search site

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/async/search_api.rst

Shows how to change the initial search settings using the configure method. The example creates a search object with site='site' and then overrides it to site='other_site'.

```pycon
>>> search = sdk.search_api.generative(site="site")
>>> search = search.configure(site="other_site")
```

--------------------------------

### Create SpeechToTextResult with one channel

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/types/speechkit.rst

Creates a SpeechToTextResult with a single channel result. This is the setup for the subsequent examples on this page.

```pycon
>>> first_result = ChannelResult(utterances=(), speaker_analysis=None, tag="222_first")
>>> result = SpeechToTextResult(
...    _sdk=None,
...    uuid='123',
...    user_request_id='234',
...    conversation_analysis=None,
...    llm_post_process_result=None,
...    unknown_events=(),
...    status_events=(),
...    channels={
...        '222_first': first_result,
...    }
... )
```

--------------------------------

### LLMPostProcessing creation and instruction examples

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/internals/bases.rst

Demonstrates creating an LLMPostProcessing object with a full model URI, a model name (which builds the URI using the configured folder_id), and a model version (default 'latest'). Then shows adding instructions via with_instruction, which returns a copy rather than mutating the object, and the resulting object representation.

```pycon
>>> llm_post_processor = LLMPostProcessing('gpt://<folder_id>/yandexgt/latest')
```

```pycon
>>> llm_post_processor = LLMPostProcessing('yandexgt')
```

```pycon
>>> llm_post_processor = LLMPostProcessing('yandexgt', model_version='latest')
```

```pycon
>>> llm_post_processor = llm_post_processor.with_instruction("Make a short review")
>>> llm_post_processor = llm_post_processor.with_instruction(
...     "What the conversation topic",
...     response_format="json"
... )
```

```pycon
>>> llm_post_processor
LLMPostProcessing(model_name='yandexgt', model_version='latest', ...)
>>> llm_post_processor
LLMPostProcessing(..., instructions=(LLMPostProcessingInstruction(...), LLMPostProcessingInstruction(...)))
```

--------------------------------

### Configure by_image search settings

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/async/search_api.rst

Shows how to create a by_image search object with initial settings and then override them using the configure method. The example appears in the documentation for the run, run_from_url, and run_from_id methods.

```pycon
>>> search = sdk.search_api.by_image(site="ya.ru")
>>> search = search.configure(site="yandex.ru")
```

--------------------------------

### setup_default_logging

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/async/sdk.rst

Sets up the default logging configuration for the SDK.

```APIDOC
## setup_default_logging

### Description
Sets up the default logging configuration.

### Method
`setup_default_logging(log_level='INFO', log_format='[%(levelname)1.1s %(asctime)s %(name)s:%(lineno)d] %(message)s', date_format='%Y-%m-%d %H:%M:%S')`

### Parameters
- **log_level** (str) – Logging level (default: 'INFO')
- **log_format** (str) – Logging format string (default: '[%(levelname)1.1s %(asctime)s %(name)s:%(lineno)d] %(message)s')
- **date_format** (str) – Date format string (default: '%Y-%m-%d %H:%M:%S')

### Notes
Read more about log_levels, log_format, and date_format in Python documentation (logging).
```

--------------------------------

### LangFilterType example

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/types/search_api.rst

Example of a language filter dictionary for generative search. The required key 'lang' is a string, e.g., 'ru'.

```pycon
>>> filter_ = {'lang': 'ru'}
```

--------------------------------

### setup_default_logging(log_level='INFO', log_format='[%(levelname)1.1s %(asctime)s %(name)s:%(lineno)d] %(message)s', date_format='%Y-%m-%d %H:%M:%S')

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/async/sdk.rst

Function to set up default logging configuration for the SDK.

```APIDOC
## setup_default_logging(log_level='INFO', log_format='[%(levelname)1.1s %(asctime)s %(name)s:%(lineno)d] %(message)s', date_format='%Y-%m-%d %H:%M:%S')

### Description
Configures default logging for the SDK with specified log level, format, and date format.

### Parameters
- **log_level** (str) - Optional - Logging level, default is 'INFO'.
- **log_format** (str) - Optional - Format string for log messages.
- **date_format** (str) - Optional - Format string for dates in log messages.
```

--------------------------------

### FormatFilterType example

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/types/search_api.rst

Example of a format filter dictionary for generative search. The required key 'format' is a string, e.g., 'xlsx'.

```pycon
>>> filter_ = {'format': 'xlsx'}
```

--------------------------------

### DateFilterType example

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/types/search_api.rst

Example of a date filter dictionary for generative search. The required key 'date' is a string, e.g., '>20240125'.

```pycon
>>> filter_ = {'date': '>20240125'}
```

--------------------------------

### Initialize AsyncAIStudio SDK with optional auth

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/examples/async/speechkit/text_to_speech/simple_run.ipynb

Creates the AsyncAIStudio SDK instance with optional folder_id and auth arguments, and enables default logging. Authentication can alternatively be provided via environment variables.

```python
# You can set authentication using environment variables instead of the 'auth' argument:
# YC_OAUTH_TOKEN, YC_TOKEN, YC_IAM_TOKEN, or YC_API_KEY
# You can also set 'folder_id' using the YC_FOLDER_ID environment variable
sdk = AsyncAIStudio(
    # folder_id="<YC_FOLDER_ID>",
    # auth="<YC_API_KEY/YC_IAM_TOKEN>",
).setup_default_logging()
```

--------------------------------

### confluence subcommand examples

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/src/yandex_ai_studio_sdk/cli/search_index/README.md

Examples of the confluence subcommand: public pages, multiple pages, private instance with environment variables, and exporting as HTML.

```bash
# Public Confluence page (no auth)
yandex-ai-studio vector-stores confluence \
  --page-url "https://cwiki.apache.org/confluence/pages/viewpage.action?pageId=34840000"

# Multiple pages
yandex-ai-studio vector-stores confluence \
  --page-url "https://your-domain.atlassian.net/wiki/spaces/ENG/pages/111/Overview" \
  --page-url "https://your-domain.atlassian.net/wiki/spaces/ENG/pages/222/Architecture"

# Private instance with environment variables
export CONFLUENCE_USERNAME=alice@example.com
export CONFLUENCE_API_TOKEN=ATATT3xFf...xxx
yandex-ai-studio vector-stores confluence \
  --page-url "https://your-domain.atlassian.net/wiki/spaces/ENG/pages/123456/Design"

# Export as HTML
yandex-ai-studio vector-stores confluence \
  --page-url "https://your-domain.atlassian.net/wiki/spaces/ENG/pages/123456/Design" \
  --export-format html
```

--------------------------------

### LLMPostProcessing.__init__

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/async/speechkit/speech_to_text.rst

Constructor for the LLMPostProcessing class.

```APIDOC
## LLMPostProcessing.__init__

### Description
Initializes a new LLMPostProcessing object with the specified model name, model version, and optional instructions.

### Method
Constructor

### Parameters
- **model_name** (str) - Required - The model name or full URI. If the name contains '://', it is treated as a full URI. Otherwise constructs a URI in the form 'gpt://<folder_id>/<model_name>/<model_version>'.
- **model_version** (str) - Optional - The version of the model to use. Defaults to 'latest'.
- **instructions** (Sequence[LLMPostProcessingInstruction]) - Optional - A list of instructions to perform with transcription. Defaults to ().

### Return Type
None
```

--------------------------------

### Run Sphinx autobuild for live preview

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/CONTRIBUTING.md

Run a local webserver that automatically rebuilds documentation when files change. Available at http://localhost:8000 by default.

```bash
sphinx-autobuild docs docs/_build --watch src
```

--------------------------------

### get

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/sync/files.rst

Retrieves a file by its ID.

```APIDOC
## get(file_id, timeout=60)

### Description
Retrieves a file by its ID.

### Method
get

### Parameters
- **file_id** (str) - Required - The unique identifier of the file to retrieve.
- **timeout** (float) - Optional - Timeout for the operation in seconds. Defaults to 60.

### Return Type
File
```

--------------------------------

### Initialize AsyncAIStudio and create yandex-art model

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/examples/async/image_generation/run_deferred.ipynb

Initializes the AsyncAIStudio SDK and creates an image generation model for 'yandex-art'. Authentication can be provided via environment variables (YC_OAUTH_TOKEN, YC_TOKEN, YC_IAM_TOKEN, YC_API_KEY) or the 'auth' argument; folder_id can be set via YC_FOLDER_ID.

```python
from yandex_ai_studio_sdk import AsyncAIStudio
# You can set authentication using environment variables instead of the 'auth' argument:
# YC_OAUTH_TOKEN, YC_TOKEN, YC_IAM_TOKEN, or YC_API_KEY
# You can also set 'folder_id' using the YC_FOLDER_ID environment variable
sdk = AsyncAIStudio(
    # folder_id="<YC_FOLDER_ID>",
    # auth="<YC_API_KEY/YC_IAM_TOKEN>",
)
model = sdk.models.image_generation('yandex-art')
```

--------------------------------

### LLMPostProcessing constructor examples

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/async/speechkit/speech_to_text.rst

Shows three ways to create an LLMPostProcessing object: with a full model URI, with a model name (URI built from configured folder_id), and with an explicit model_version. The model_version defaults to 'latest'.

```pycon
>>> llm_post_processor = LLMPostProcessing('gpt://<folder_id>/yandexgt/latest')
```

```pycon
>>> llm_post_processor = LLMPostProcessing('yandexgt')
```

```pycon
>>> llm_post_processor = LLMPostProcessing('yandexgt', model_version='latest')
```

--------------------------------

### AsyncFiles.get()

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/async/sdk.rst

Method to get a file.

```APIDOC
## AsyncFiles.get()

### Description
Retrieves a file.

### Method
get

### Parameters
None
```

--------------------------------

### s3 subcommand usage examples

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/src/yandex_ai_studio_sdk/cli/search_index/README.md

Four usage examples for the s3 subcommand: indexing an entire bucket, filtering by prefix, including only PDF files, and using Yandex Object Storage with a custom endpoint and region.

```bash
# Index entire bucket
yandex-ai-studio vector-stores s3 my-bucket

# Index only a specific prefix
yandex-ai-studio vector-stores s3 my-bucket --prefix docs/

# Include only PDF files
yandex-ai-studio vector-stores s3 my-bucket --include-pattern "*.pdf"

# Use Yandex Object Storage
yandex-ai-studio vector-stores s3 my-bucket \
  --endpoint-url https://storage.yandexcloud.net \
  --region-name ru-central1
```

--------------------------------

### configure(site=Undefined, host=Undefined, url=Undefined, fix_misspell=Undefined, enable_nrfm_docs=Undefined, search_filters=Undefined)

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/async/search_api.rst

Returns a new object with config fields overridden by passed values. Parameters site, host, and url are mutually exclusive.

```APIDOC
## configure(site=Undefined, host=Undefined, url=Undefined, fix_misspell=Undefined, enable_nrfm_docs=Undefined, search_filters=Undefined)

### Description
Returns the new object with config fields overrode by passed values. To learn more about parameters and their formats and possible values, refer to generative search documentation.

### Method
configure

### Parameters
- **site** (str | Sequence[str] | Undefined | None) - Optional - Parameter for limiting search to specific location or list of sites.
- **host** (str | Sequence[str] | Undefined | None) - Optional - Parameter for limiting search to specific location or list of hosts.
- **url** (str | Sequence[str] | Undefined | None) - Optional - Parameter for limiting search to specific location or list of URLs.
- **fix_misspell** (bool | Undefined | None) - Optional - Tells backend to fix or not to fix misspellings in queries.
- **enable_nrfm_docs** (bool | Undefined | None) - Optional - Tells backend to include or not to include pages not available via direct clicks from given sites/hosts/urls to search result.
- **search_filters** (Sequence[DateFilterType | FormatFilterType | LangFilterType] | DateFilterType | FormatFilterType | LangFilterType | Undefined | None) - Optional - Allows to limit search results with additional filters.

### Return Type
Self
```

--------------------------------

### AsyncFile.get_url()

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/async/sdk.rst

Method to get the URL of a file.

```APIDOC
## AsyncFile.get_url()

### Description
Gets the URL of the file.

### Method
get_url

### Parameters
None
```

--------------------------------

### Initialize AIStudio SDK with authentication

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/examples/sync/speechkit/text_to_speech/simple_run.ipynb

Creates an AIStudio instance. Authentication can be set via environment variables (YC_OAUTH_TOKEN, YC_TOKEN, YC_IAM_TOKEN, YC_API_KEY) and folder ID via YC_FOLDER_ID, or by passing arguments directly. Enables default logging.

```python
# You can set authentication using environment variables instead of the 'auth' argument:
# YC_OAUTH_TOKEN, YC_TOKEN, YC_IAM_TOKEN, or YC_API_KEY
# You can also set 'folder_id' using the YC_FOLDER_ID environment variable
sdk = AIStudio(
    # folder_id="<YC_FOLDER_ID>",
    # auth="<YC_API_KEY/YC_IAM_TOKEN>",
).setup_default_logging()
```

--------------------------------

### AsyncTuning.get()

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/async/sdk.rst

Method to get a tuning operation.

```APIDOC
## AsyncTuning.get()

### Description
Retrieves a tuning operation.

### Method
get

### Parameters
None
```

--------------------------------

### get

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/async/search_indexes.rst

Retrieves a search index by its unique identifier.

```APIDOC
## get

### Description
Retrieves a search index by its unique identifier.

### Method
async get

### Parameters
- **search_index_id** (str) - Required - The unique identifier of the search index to retrieve.
- **timeout** (float) - Optional - The time to wait for the operation to complete. Defaults to 60 seconds.

### Returns
AsyncSearchIndex
```

--------------------------------

### Slice channels

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/types/speechkit.rst

Uses slicing to get a tuple of channels in sorted order.

```pycon
>>> assert result[0:2] == (second_result, first_result)
```

--------------------------------

### SpeechToTextResult

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/types/speechkit.rst

Documentation for the SpeechToTextResult class, including its constructor, properties, methods, and usage examples.

```APIDOC
## SpeechToTextResult

### Description
A class representing the result of a speech recognition request. It contains one or more channel results, and provides proxy-properties to access the first channel's data when only one channel is present.

### Class
`yandex_ai_studio_sdk._speechkit.speech_to_text.result.top_containers.SpeechToTextResult`

### Constructor Parameters
- **`_sdk`** (object) - Required - Internal SDK reference.
- **`uuid`** (str) - Required - Internal session identifier.
- **`user_request_id`** (str) - Required - User session identifier.
- **`conversation_analysis`** (ConversationAnalysis | None) - Required - Conversation statistics.
- **`llm_post_process_result`** (LLMPostProcessResult | None) - Required - Result of LLM post-processing (also known as Summarization).
- **`unknown_events`** (tuple[SpeechToTextStreamingEvent, ...]) - Required - Unknown streaming events.
- **`status_events`** (tuple[SpeechToTextStreamingEvent, ...]) - Required - Status streaming events.
- **`channels`** (dict[str, ChannelResult]) - Required - Mapping of channel tags to channel results.

### Properties
- **`text`** (str) - Returns the final text of the first channel. Raises if more than one channel.
- **`final_text`** (str) - Returns the final text of the first channel. Raises if more than one channel.
- **`final_refinement_text`** (str | None) - Returns the final refinement text of the first channel. Raises if more than one channel.
- **`speaker_analysis`** (SpeakerAnalysis | None) - Returns speaker analysis for the first channel. Raises if more than one channel.

### Methods
#### `get_channel(sequential_number)`
Get channel by its sequential number among all channels, sorted by tags.

**Parameters:**
- **`sequential_number`** (int) - Required - The sequential index (can be negative).

**Return type:**
- `ChannelResult`

#### `count(value)`
Return number of occurrences of value.

#### `index(value)`
Return first index of value. Raises ValueError if not present.

### Attributes
- **`role`** (ClassVar[str]) - Always `'user'`. Required for TextMessage Protocol compatibility.
- **`uuid`** (str) - Internal session identifier.
- **`user_request_id`** (str) - User session identifier.
- **`conversation_analysis`** (ConversationAnalysis | None) - Conversation statistics.
- **`llm_post_process_result`** (LLMPostProcessResult | None) - Result of LLM post-processing.
- **`unknown_events`** (tuple[SpeechToTextStreamingEvent, ...]) - Unknown streaming events.
- **`status_events`** (tuple[SpeechToTextStreamingEvent, ...]) - Status streaming events.
- **`channels`** (dict[str, ChannelResult]) - Mapping of channel tags to channel results.

### Usage Examples

#### Accessing proxy-properties (single channel)
```pycon
>>> result = SpeechToTextResult(
...    _sdk=None,
...    uuid='123',
...    user_request_id='234',
...    conversation_analysis=None,
...    llm_post_process_result=None,
...    unknown_events=(),
...    status_events=(),
...    channels={'222_first': first_result}
... )
>>> text = result.text
>>> final_text = result.text
```

#### Accessing channels by tag
```pycon
>>> assert result['222_first'] == result.channels['222_first'] is first_result
```

#### Accessing channels by index (sorted by tags)
```pycon
>>> assert result[0] == result.get_channel(0) is second_result
>>> assert result[-1] == result.get_channel(-1) is first_result
```

#### Slicing and iteration
```pycon
>>> assert result[0:2] == (second_result, first_result)
>>> assert list(result) == [second_result, first_result]
```

### Error Handling
- Accessing proxy-properties (`text`, `final_text`, `final_refinement_text`, `speaker_analysis`) when more than one channel is present raises a `ValueError`.
- `index()` raises `ValueError` if the value is not present.
```

--------------------------------

### Adding instructions to LLMPostProcessing

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/docs/async/speechkit/speech_to_text.rst

Demonstrates adding instructions to an LLMPostProcessing object using .with_instruction, which returns a copy rather than mutating the original. The second call shows passing a response_format of 'json'. The output shows the resulting object with instructions.

```pycon
>>> llm_post_processor = llm_post_processor.with_instruction("Make a short review")
>>> llm_post_processor = llm_post_processor.with_instruction(
...     "What the conversation topic",
...     response_format="json"
... )
```

```pycon
>>> llm_post_processor
LLMPostProcessing(model_name='yandexgt', model_version='latest', ...)
>>> llm_post_processor
LLMPostProcessing(..., instructions=(LLMPostProcessingInstruction(...), LLMPostProcessingInstruction(...)))
```

--------------------------------

### Create and configure text-to-speech object

Source: https://github.com/yandex-cloud/yandex-ai-studio-sdk/blob/master/examples/sync/speechkit/text_to_speech/simple_run.ipynb

Creates a TTS object with voice 'kirill' and audio format PCM16 at the defined sample rate. Shows how to reconfigure the object after creation, including alternative audio format specification, loudness normalization to LUFS, and speed setting.

```python
tts = sdk.speechkit.text_to_speech(
    voice='kirill',
    audio_format=f'PCM16({SAMPLERATE})',
)   
# you can configure TTS object after creation
tts = tts.configure(
    # alternative way to specify audio_format
    audio_format=tts.AudioFormat.PCM16(SAMPLERATE),
    loudness_normalization=tts.LoudnessNormalization.LUFS,
    speed=1.5,
)   
```
Tokens:
Raw