| Name | aiaas-falcon JSON |
| Version |
0.2.2
JSON |
| download |
| home_page | |
| Summary | This python package help to interact with Generative AI - Large Language Models. It interacts with AIaaS LLM , AIaaS embedding , AIaaS Audio set of APIs to cater the request. |
| upload_time | 2024-01-04 09:39:33 |
| maintainer | |
| docs_url | None |
| author | Your Name |
| requires_python | >=3.8.1,<4.0.0 |
| license | MIT |
| keywords |
|
| VCS |
|
| bugtrack_url |
|
| requirements |
No requirements were recorded.
|
| Travis-CI |
No Travis.
|
| coveralls test coverage |
No coveralls.
|
# AIaaS Falcon
<h4 align="center">
<p>
<a href="#shield-installation">Installation</a> |
<a href="#fire-quickstart">Quickstart</a> |
<p>
</h4>
## Description
AIaaS_Falcon is Generative AI - LLM library interacts with different model api endpoints., allowing operations such as listing models, creating embeddings, and generating text based on certain configurations. AIaaS_Falcon helps to invoking the RAG pipeline in seconds.
## Supported Endpoint Types:
- Azure OpenAI
- SingtelGPT
- Dev_Quantized
- Dev_Full
## :shield: Installation
Ensure you have the `requests` and `google-api-core` libraries installed:
```bash
pip install aiaas-falcon
```
if you want to install from source
```bash
git clone https://github.com/Praveengovianalytics/AIaaS_falcon && cd AIaaS_falcon
pip install -e .
```
### Methods
### `Falcon` Class
- `__init__ (config)`
Intialise the Falcon object with endpoint configs. \
Parameter:
- api_key : API Key
- api_name: Name for endpoint
- api_endpoint: Type of endpoint ( can be azure, dev_quan, dev_full, prod)
- host_name_port: Host and Port Inforation
- use_pil: Activate Personal Identifier Information Limit Protection (Boolean)
- protocol: HTTP/ HTTPS
- api_type: Subroute if needed
- use_pii: Whether current endpoint need Personal Identifier Information Limit Protection
- log_key: Auth Key to use the Application
- `current_active()`
Check current endpoint active
- `add_endpoint(api_name,protocol,host_name_port,api_endpoint,api_key,use_pil=False)`
Add a new endpoint. \
Parameter:
- api_key : API Key
- api_name: Name for endpoint
- api_endpoint: Type of endpoint ( can be azure, dev_quan, dev_full, prod)
- host_name_port: Host and Port Inforation
- use_pii: Activate Personal Identifier Information Limit Protection (Boolean)
- protocol: HTTP/ HTTPS
- use_pil: Whether current endpoint need Personal Identifier Information Limit Protection
- `list_endpoint()`
List out all endpoints in the endpoint management manager
- `set_endpoint(name)`
Set target endpoint to active \
Parameter:
- name : Target endpoint's name
- `remove_endpoint(name)`
Delete endpoint by name \
Parameter:
- name : Target endpoint's name
- `current_pii()`
Check current Personal Identifier Information Protection activation status
- `switch_pii()`
Switch current Personal Identifier Information Protection activation status
- `list_models()`
List out models available
- `initalise_pii()`
Download and intialise PII Protection. \
Note: This does not activate PII but initialise dependencies
- `health()`
Check health of current endpoint
- `create_embedding(file_path)`
Create embeddings by sending files to the API. \
Parameter:
- file_path: Path to file
- `generate_text_full(query="",
context="",
use_file=0,
model="",
chat_history=[],
max_new_tokens: int = 200,
temperature: float = 0,
top_k: int = -1,
frequency_penalty: int = 0,
repetition_penalty: int = 1,
presence_penalty: float = 0,
fetch_k=100000,
select_k=4,
api_version='2023-05-15',
guardrail={'jailbreak': False, 'moderation': False},
custom_guardrail=None)` \
Generate text using LLM endpoint. Note: Some parameter of the endpoint is endpoint-specific. \
Parameter:
- query: a string of your prompt
- use_file: Whether to take file to context in generation. Only applies to dev_full and dev_quan. Need to `create_embedding` before use.
- model: a string on the model to use. You can use ` list_models` to check for model available.
- chat_history: an array of chat history between user and bot. Only applies to dev_full and dev_quan. (Beta)
- max_new_token: maximum new token to generate. Must be integer.
- temperature: Float that controls the randomness of the sampling. Lower
values make the model more deterministic, while higher values make
the model more random. Zero means greedy sampling.
- top_k: Integer that controls the number of top tokens to consider.
- frequency_penalty: Float that penalizes new tokens based on their
frequency in the generated text so far.
- repetition_penalty: Float that penalizes new tokens based on whether
they appear in the prompt and the generated text so far.
- presence_penalty: Float that penalizes new tokens based on whether they
appear in the generated text so far
- fetch_k: Use for document retrival. Include how many element in searching. Only applies when `use_file` is 1
- select k: Use to select number of document for document retrieval. Only applies when `use_file` is 1
- api_version: Only applies for azure endpoint
- guardrail: Whether to use the default jailbreak guardrail and moderation guardrail
- custom_guardrail: Path to custom guardrail .yaml file. The format can be found in sample.yaml
- ` evaluate_parameter(config)`
Carry out grid search for parameter \
Parameter:
- config: A dict. The dict must contain model and query. Parameter to grid search must be a list.
- model: a string of model
- query: a string of query
- **other parameter (eg: "temperature":list(np.arange(0,2,0.5))
- `decrypt_hash(encrypted_data)`
Decret the configuration from experiment id.
Parameter:
- encrypted_data: a string of id
## :fire: Quickstart
```python
from aiaas_falcon import Falcon
model=Falcon(api_name="azure_1",protocol='https',host_name_port='example.com',api_key='API_KEY',api_endpoint='azure',log_key="KEY")
model.list_models()
model.generate_text_full(query="Hello, introduce yourself",model='gpt-35-turbo-0613-vanilla',api_version='2023-05-15')
```
## Conclusion
AIaaS_Falcon library simplifies interactions with the LLM API's, providing a straightforward way to perform various operations such as listing models, creating embeddings, and generating text.
## Authors
- [@Praveengovianalytics](https://github.com/Praveengovianalytics)
- [@zhuofan](https://github.com/zhuofan-16)
## Google Colab
- [Get start with aiaas_falcon](https://colab.research.google.com/drive/1haZ-1fD4htQuNF2zzyrUSTP90KRls1dC?usp=sharing)
## Badges
[](https://choosealicense.com/licenses/mit/)
Raw data
{
"_id": null,
"home_page": "",
"name": "aiaas-falcon",
"maintainer": "",
"docs_url": null,
"requires_python": ">=3.8.1,<4.0.0",
"maintainer_email": "",
"keywords": "",
"author": "Your Name",
"author_email": "you@example.com",
"download_url": "https://files.pythonhosted.org/packages/e0/03/a1476698ce328157bbe6ee5a48d4aa0b09c2be24e7a79a172ec8d7541bc1/aiaas_falcon-0.2.2.tar.gz",
"platform": null,
"description": "\n\n# AIaaS Falcon\n\n\n<h4 align=\"center\">\n <p>\n <a href=\"#shield-installation\">Installation</a> |\n <a href=\"#fire-quickstart\">Quickstart</a> |\n <p>\n</h4>\n\n\n\n\n## Description\n\nAIaaS_Falcon is Generative AI - LLM library interacts with different model api endpoints., allowing operations such as listing models, creating embeddings, and generating text based on certain configurations. AIaaS_Falcon helps to invoking the RAG pipeline in seconds. \n\n## Supported Endpoint Types:\n- Azure OpenAI\n- SingtelGPT\n- Dev_Quantized\n- Dev_Full\n\n## :shield: Installation\n\nEnsure you have the `requests` and `google-api-core` libraries installed:\n\n```bash\npip install aiaas-falcon\n```\n\n\nif you want to install from source\n\n```bash\ngit clone https://github.com/Praveengovianalytics/AIaaS_falcon && cd AIaaS_falcon\npip install -e .\n```\n\n### Methods\n### `Falcon` Class\n- `__init__ (config)`\nIntialise the Falcon object with endpoint configs. \\\nParameter: \n - api_key : API Key\n - api_name: Name for endpoint\n - api_endpoint: Type of endpoint ( can be azure, dev_quan, dev_full, prod)\n - host_name_port: Host and Port Inforation\n - use_pil: Activate Personal Identifier Information Limit Protection (Boolean)\n - protocol: HTTP/ HTTPS\n - api_type: Subroute if needed\n - use_pii: Whether current endpoint need Personal Identifier Information Limit Protection\n - log_key: Auth Key to use the Application\n- `current_active()`\nCheck current endpoint active\n- `add_endpoint(api_name,protocol,host_name_port,api_endpoint,api_key,use_pil=False)`\nAdd a new endpoint. \\\nParameter:\n - api_key : API Key\n - api_name: Name for endpoint\n - api_endpoint: Type of endpoint ( can be azure, dev_quan, dev_full, prod)\n - host_name_port: Host and Port Inforation\n - use_pii: Activate Personal Identifier Information Limit Protection (Boolean)\n - protocol: HTTP/ HTTPS\n - use_pil: Whether current endpoint need Personal Identifier Information Limit Protection\n- `list_endpoint()`\nList out all endpoints in the endpoint management manager\n- `set_endpoint(name)`\nSet target endpoint to active \\\nParameter:\n - name : Target endpoint's name\n \n- `remove_endpoint(name)`\nDelete endpoint by name \\\nParameter:\n - name : Target endpoint's name\n\n- `current_pii()`\nCheck current Personal Identifier Information Protection activation status\n\n- `switch_pii()`\nSwitch current Personal Identifier Information Protection activation status\n- `list_models()`\nList out models available\n- `initalise_pii()`\nDownload and intialise PII Protection. \\\nNote: This does not activate PII but initialise dependencies\n\n- `health()`\nCheck health of current endpoint\n\n- `create_embedding(file_path)`\nCreate embeddings by sending files to the API. \\\nParameter:\n - file_path: Path to file \n\n- `generate_text_full(query=\"\",\n context=\"\",\n use_file=0,\n model=\"\",\n chat_history=[],\n max_new_tokens: int = 200,\n temperature: float = 0,\n top_k: int = -1,\n frequency_penalty: int = 0,\n repetition_penalty: int = 1,\n presence_penalty: float = 0,\n fetch_k=100000,\n select_k=4,\n api_version='2023-05-15',\n guardrail={'jailbreak': False, 'moderation': False},\n custom_guardrail=None)` \\\n Generate text using LLM endpoint. Note: Some parameter of the endpoint is endpoint-specific. \\\n Parameter: \n - query: a string of your prompt\n - use_file: Whether to take file to context in generation. Only applies to dev_full and dev_quan. Need to `create_embedding` before use.\n - model: a string on the model to use. You can use ` list_models` to check for model available.\n - chat_history: an array of chat history between user and bot. Only applies to dev_full and dev_quan. (Beta)\n - max_new_token: maximum new token to generate. Must be integer.\n - temperature: Float that controls the randomness of the sampling. Lower\n values make the model more deterministic, while higher values make\n the model more random. Zero means greedy sampling.\n - top_k: Integer that controls the number of top tokens to consider.\n - frequency_penalty: Float that penalizes new tokens based on their\n frequency in the generated text so far.\n - repetition_penalty: Float that penalizes new tokens based on whether\n they appear in the prompt and the generated text so far.\n - presence_penalty: Float that penalizes new tokens based on whether they\n appear in the generated text so far\n - fetch_k: Use for document retrival. Include how many element in searching. Only applies when `use_file` is 1\n - select k: Use to select number of document for document retrieval. Only applies when `use_file` is 1\n - api_version: Only applies for azure endpoint\n - guardrail: Whether to use the default jailbreak guardrail and moderation guardrail\n - custom_guardrail: Path to custom guardrail .yaml file. The format can be found in sample.yaml\n \n- ` evaluate_parameter(config)`\nCarry out grid search for parameter \\\nParameter:\n - config: A dict. The dict must contain model and query. Parameter to grid search must be a list. \n - model: a string of model\n - query: a string of query\n - **other parameter (eg: \"temperature\":list(np.arange(0,2,0.5))\n- `decrypt_hash(encrypted_data)`\nDecret the configuration from experiment id.\nParameter:\n - encrypted_data: a string of id\n\n\n## :fire: Quickstart\n\n```python\nfrom aiaas_falcon import Falcon\nmodel=Falcon(api_name=\"azure_1\",protocol='https',host_name_port='example.com',api_key='API_KEY',api_endpoint='azure',log_key=\"KEY\")\nmodel.list_models()\nmodel.generate_text_full(query=\"Hello, introduce yourself\",model='gpt-35-turbo-0613-vanilla',api_version='2023-05-15')\n\n```\n\n\n\n## Conclusion\n\nAIaaS_Falcon library simplifies interactions with the LLM API's, providing a straightforward way to perform various operations such as listing models, creating embeddings, and generating text.\n\n## Authors\n\n- [@Praveengovianalytics](https://github.com/Praveengovianalytics)\n- [@zhuofan](https://github.com/zhuofan-16)\n\n## Google Colab\n\n- [Get start with aiaas_falcon](https://colab.research.google.com/drive/1haZ-1fD4htQuNF2zzyrUSTP90KRls1dC?usp=sharing)\n\n## Badges\n\n[](https://choosealicense.com/licenses/mit/)\n",
"bugtrack_url": null,
"license": "MIT",
"summary": "This python package help to interact with Generative AI - Large Language Models. It interacts with AIaaS LLM , AIaaS embedding , AIaaS Audio set of APIs to cater the request.",
"version": "0.2.2",
"project_urls": null,
"split_keywords": [],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "4650ddc2eb44982c37d3ee21e32da0f89bff76a1214c69668dfc3882dab5f8ee",
"md5": "3062e1f44241d1becf1e594b7608ec3c",
"sha256": "2e9a4f2db99b39e99f70f70e1673757f443398da7b1b04139d72b92267c240be"
},
"downloads": -1,
"filename": "aiaas_falcon-0.2.2-py3-none-any.whl",
"has_sig": false,
"md5_digest": "3062e1f44241d1becf1e594b7608ec3c",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": ">=3.8.1,<4.0.0",
"size": 6229,
"upload_time": "2024-01-04T09:39:32",
"upload_time_iso_8601": "2024-01-04T09:39:32.475846Z",
"url": "https://files.pythonhosted.org/packages/46/50/ddc2eb44982c37d3ee21e32da0f89bff76a1214c69668dfc3882dab5f8ee/aiaas_falcon-0.2.2-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": "",
"digests": {
"blake2b_256": "e003a1476698ce328157bbe6ee5a48d4aa0b09c2be24e7a79a172ec8d7541bc1",
"md5": "75482c7395ff4711c5a9ae02e81d0dd4",
"sha256": "5bd74805c884de48d93c8aa4ef0d0831daffac9888a62309c358f3c20988c06b"
},
"downloads": -1,
"filename": "aiaas_falcon-0.2.2.tar.gz",
"has_sig": false,
"md5_digest": "75482c7395ff4711c5a9ae02e81d0dd4",
"packagetype": "sdist",
"python_version": "source",
"requires_python": ">=3.8.1,<4.0.0",
"size": 5619,
"upload_time": "2024-01-04T09:39:33",
"upload_time_iso_8601": "2024-01-04T09:39:33.773743Z",
"url": "https://files.pythonhosted.org/packages/e0/03/a1476698ce328157bbe6ee5a48d4aa0b09c2be24e7a79a172ec8d7541bc1/aiaas_falcon-0.2.2.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2024-01-04 09:39:33",
"github": false,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"lcname": "aiaas-falcon"
}
