LibreOpenAI 0.1.6

Libre-OpenAI (Beta)

An Open AI Nuget for .NET applications (.NET 8 or higher), It is "Libre" (free and for free). Remember, Libre-OpenAI and its Extensions are Beta versions, they are under construction!

You can download our Nuget from:

https://www.nuget.org/packages/LibreOpenAI/

Please, see these unit tests to learn how to use this Nuget:

https://github.com/maxiunlm/Libre-OpenAI/blob/main/LibreOpenAIUnitTestProject/OpenAiUnitTest.cs

And see the Libre OpenAI Extensions Nuget to learn how to use the extensions of the Libre Open AI Nuget:

https://github.com/maxiunlm/Libre-OpenAI/blob/main/LibreOpenAIExtensions/README.md

Support this project 💖

If you find this project useful, consider supporting us by making a donation via PayPal. Your support helps keep this project alive and maintained.

• Donate to Libre-OpenAI (Beta):

Donate via QR Code

You can also scan the QR code below to make a donation:

Settings & Setup

Before you can use the OpenAI API, you need to set up an account and create an API key. You can do this by following the instructions on the OpenAI website. Unce you have your API key, you can use it to authenticate your requests to the OpenAI API. For that, you need to set the API key in the OPENAI_API_KEY or LIBRE_OPEN_AI_API_KEY environment variables, they are th same for this Nuget.

• OPENAI_API_KEY is most standard environment variable name for OpenAI API key.
• LIBRE_OPEN_AI_API_KEY is the environment variable name for Libre OpenAI API key. This environment variable lets you using OPENAI_API_KEY for other APIs or Nugets or even if you need to test more than one API KEY for each.

OpenAI API Key Setup

However, this little guide will help you to set up your API key in a few simple steps.

1. Create an OpenAI Account

To generate an API key, you first need to create an OpenAI account:

• Go to OpenAI Sign-Up.
• Fill out the registration form with your email and password.
• Verify your email address by following the instructions in your inbox.

2. Generate an API Key

Once your account is set up, follow these steps to create an API key:

1. Log in to OpenAI Platform.
2. Click on your profile avatar in the top-right corner.
3. Select "View API Keys".
4. Click "Create new secret key".
5. Copy and store the generated key securely (it will only be displayed once).

3. OpenAI Documentation

For more details, visit the official OpenAI API documentation:
OpenAI API Docs

How to use Libre Open AI

Create Completions

Using IRequestBody

    IRequestBody request = new RequestBody
    {
        Model = defaultModel,
        MaxCompletionTokens = defaultMaxCompletionTokens,
        Messages = new List<MessageRequest> {
            new MessageRequest {
                Role = defaultRole,
                Content = "What is the capital of France?" // NOTE: complete here your request question
            }
        }
    };
    IOpenAI openAi = new OpenAI();

    IChatCompletionResponse result = await openAi.Chat.Completions.Create(request);
    string contentResult = result.Choices.First().Message.Content;

Using JSON

    string request = "{\"max_completion_tokens\":50,\"model\":\"gpt-3.5-turbo\",\"messages\":[{\"role\":\"user\",\"content\":\"What is the capital of France?\"}],\"n\":1}";
    IOpenAI openAi = new OpenAI();

    IChatCompletionResponse result = await openAi.Chat.Completions.Create(request);
    string contentResult = result.Choices.First().Message.Content;

Using dynamic type

    dynamic request = new {
        model = defaultModel,
        max_completion_tokens = defaultMaxCompletionTokens,
        messages = new[] {
            new {
                role = defaultRole,
                content = ResponseFakes.whatIsTheCapitalOfFrance
            }
        }
    };
    IOpenAI openAi = new OpenAI();

    IChatCompletionResponse result = await openAi.Chat.Completions.Create(request);
    string contentResult = result.Choices.First().Message.Content;

Returning dynamic

Using IRequestBody

    IRequestBody request = new RequestBody
    {
        Model = defaultModel,
        MaxCompletionTokens = defaultMaxCompletionTokens,
        Messages = new List<MessageRequest> {
            new MessageRequest {
                Role = defaultRole,
                Content = "What is the capital of France?" // NOTE: complete here your request question
            }
        }
    };
    IOpenAI openAi = new OpenAI();

    dynamic result = await openAi.Chat.Completions.CreateDynamic(request);

    // result will be a dynamic object based on the JSON result
    string content = result.choices[0].message.content.Value;

    // OR, result will be a IDictionary<string, object> object based on the JSON result
    ((IDictionary<string, object>)
        ((IDictionary<string, object>)
            ((IList<object>)result["choices"])[0])["message"])["content"];

Using JSON

    string request = "{\"max_completion_tokens\":50,\"model\":\"gpt-3.5-turbo\",\"messages\":[{\"role\":\"user\",\"content\":\"What is the capital of France?\"}],\"n\":1}";
    IOpenAI openAi = new OpenAI();

    dynamic result = await openAi.Chat.Completions.CreateDynamic(request);

    // result will be a dynamic object based on the JSON result
    string content = result.choices[0].message.content.Value;

    // OR, result will be a IDictionary<string, object> object based on the JSON result
    ((IDictionary<string, object>)
        ((IDictionary<string, object>)
            ((IList<object>)result["choices"])[0])["message"])["content"];

Using dynamic type

    dynamic request = new {
        model = defaultModel,
        max_completion_tokens = defaultMaxCompletionTokens,
        messages = new[] {
            new {
                role = defaultRole,
                content = ResponseFakes.whatIsTheCapitalOfFrance
            }
        }
    };
    IOpenAI openAi = new OpenAI();

    dynamic result = await openAi.Chat.Completions.CreateDynamic(request);

    // result will be a IDictionary<string, object> object based on the JSON result
    string content = result.choices[0].message.content.Value;

    // OR, result will be a IDictionary<string, object> object based on the JSON result
    ((IDictionary<string, object>)
        ((IDictionary<string, object>)
            ((IList<object>)result["choices"])[0])["message"])["content"];

Returning JSON

Using IRequestBody

    IRequestBody request = new RequestBody
    {
        Model = defaultModel,
        MaxCompletionTokens = defaultMaxCompletionTokens,
        Messages = new List<MessageRequest> {
            new MessageRequest {
                Role = defaultRole,
                Content = "What is the capital of France?" // NOTE: complete here your request question
            }
        }
    };
    IOpenAI openAi = new OpenAI();

    string result = await openAi.Chat.Completions.CreateJson(request);  
    // result will be a JSON

Using JSON

    string request = "{\"max_completion_tokens\":50,\"model\":\"gpt-3.5-turbo\",\"messages\":[{\"role\":\"user\",\"content\":\"What is the capital of France?\"}],\"n\":1}";
    IOpenAI openAi = new OpenAI();

    string result = await openAi.Chat.Completions.CreateJson(request);  
    // result will be a JSON

Using dynamic type

    dynamic request = new {
        model = defaultModel,
        max_completion_tokens = defaultMaxCompletionTokens,
        messages = new[] {
            new {
                role = defaultRole,
                content = ResponseFakes.whatIsTheCapitalOfFrance
            }
        }
    };
    IOpenAI openAi = new OpenAI();

    string result = await openAi.Chat.Completions.CreateJson(request);    
    // result will be a JSON

Create Stream Completions

Using IRequestBody

    IRequestBody request = new RequestBody
    {
        Model = "gpt-4o",
        Messages = new List<MessageRequest> {
            new MessageRequest {
                Role = "system",
                Content = "You are a helpful assistant." // NOTE: complete here your request question
            },
            new MessageRequest {
                Role = "user",
                Content = "Hello!" // NOTE: complete here your request question
            }
        },
        Stream = true
    };

    IOpenAI openAi = new OpenAI();
    
    List<IChatCompletionChunk> result = await sut.Chat.Completions.CreateStream(request);
    // For ex.: contentResult will contain all the result prats separated by "\t" (TAB character).
    string contentResult = string.Join('\t', result.SelectMany(o => o.Choices.Select(o => o.Delta.Content)).ToList());

Using JSON

    string request = "{\"max_completion_tokens\":50,\"model\":\"gpt-4o\",\"messages\":[{\"role\":\"user\",\"content\":\"What is the capital of France?\"}],\"n\":1}";
    IOpenAI openAi = new OpenAI();
    
    List<IChatCompletionChunk> result = await sut.Chat.Completions.CreateStream(request);
    // For ex.: contentResult will contain all the result prats separated by "\t" (TAB character).
    string contentResult = string.Join('\t', result.SelectMany(o => o.Choices.Select(o => o.Delta.Content)).ToList());

Using dynamic type

    dynamic request = new
    {
        model = "gpt-4o",
        messages = new[]
            {
                new { role = "system", content = "You are a helpful assistant." },
                new { role = "user", content = "Hello!" }
            },
        stream = true
    };
    IOpenAI openAi = new OpenAI();
    
    List<IChatCompletionChunk> result = await sut.Chat.Completions.CreateStream(request);
    // For ex.: contentResult will contain all the result prats separated by "\t" (TAB character).
    string contentResult = string.Join('\t', result.SelectMany(o => o.Choices.Select(o => o.Delta.Content)).ToList());

Returning dynamic

Using IRequestBody

    IRequestBody request = new RequestBody
    {
        Model = "gpt-4o",
        Messages = new List<MessageRequest> {
            new MessageRequest {
                Role = "system",
                Content = "You are a helpful assistant." // NOTE: complete here your request question
            },
            new MessageRequest {
                Role = "user",
                Content = "Hello!" // NOTE: complete here your request question
            }
        },
        Stream = true
    };

    IOpenAI openAi = new OpenAI();
    
    dynamic result = await sut.Chat.Completions.CreateStreamDynamic(request);  
    // result will be a dynamic array of objects based on the JSON result
    string content = result[1].choices[0].delta.content.Value;

Using JSON

    string request = "{\"max_completion_tokens\":50,\"model\":\"gpt-4o\",\"messages\":[{\"role\":\"user\",\"content\":\"What is the capital of France?\"}],\"n\":1}";
    IOpenAI openAi = new OpenAI();
    
    dynamic result = await sut.Chat.Completions.CreateStreamDynamic(request);
    // result will be a dynamic array of objects based on the JSON result
    string content = result[1].choices[0].delta.content.Value;

Using dynamic type

    dynamic request = new
    {
        model = "gpt-4o",
        messages = new[]
            {
                new { role = "system", content = "You are a helpful assistant." },
                new { role = "user", content = "Hello!" }
            },
        stream = true
    };
    IOpenAI openAi = new OpenAI();
    
    dynamic result = await sut.Chat.Completions.CreateStreamDynamic(request);
    // result will be a dynamic array of objects based on the JSON result
    string content = result[1].choices[0].delta.content.Value;

Returning JSON

Using IRequestBody

    IRequestBody request = new RequestBody
    {
        Model = "gpt-4o",
        Messages = new List<MessageRequest> {
            new MessageRequest {
                Role = "system",
                Content = "You are a helpful assistant." // NOTE: complete here your request question
            },
            new MessageRequest {
                Role = "user",
                Content = "Hello!" // NOTE: complete here your request question
            }
        },
        Stream = true
    };

    IOpenAI openAi = new OpenAI();
    
    string result = await sut.Chat.Completions.CreateStreamJson(request);
    // result will be a JSON array

Using JSON

    string request = "{\"max_completion_tokens\":50,\"model\":\"gpt-4o\",\"messages\":[{\"role\":\"user\",\"content\":\"What is the capital of France?\"}],\"n\":1}";
    IOpenAI openAi = new OpenAI();
    
    string result = await sut.Chat.Completions.CreateStreamJson(request);
    // result will be a JSON array

Using dynamic type

    dynamic request = new
    {
        model = "gpt-4o",
        messages = new[]
            {
                new { role = "system", content = "You are a helpful assistant." },
                new { role = "user", content = "Hello!" }
            },
        stream = true
    };
    IOpenAI openAi = new OpenAI();
    
    string result = await sut.Chat.Completions.CreateStreamJson(request);
    // result will be a JSON array

Expecting raw JSON

    string request = "{\"max_completion_tokens\":50,\"model\":\"gpt-4o\",\"messages\":[{\"role\":\"user\",\"content\":\"What is the capital of France?\"}],\"n\":1}";
    IOpenAI openAi = new OpenAI();
    
    string result = await sut.Chat.Completions.CreateStreamJson(request, true);

    // result will contain a list of "data: {...}" JSON objects that can't be parsed directly as JSON !

Batches

Create

Returning dynamic

Using JSON

    string request = @"{
        ""input_file_id"": ""file-abc123"",
        ""endpoint"": ""/v1/chat/completions"",
        ""completion_window"": ""24h""
    }";
    IOpenAI openAi = new OpenAI();

    dynamic result = await openAi.Batches.CreateDynamic(request);
    // result will be a dynamic object based on the JSON result
    string id = result.id.Value;

Using dynamic type

    dynamic request = new {
        input_file_id = "file-abc123",
        endpoint = "/v1/chat/completions",
        completion_window = "24h"
    };
    IOpenAI openAi = new OpenAI();

    dynamic result = await openAi.Batches.CreateDynamic(request);
    // result will be a dynamic object based on the JSON result
    string id = result.id.Value;

Returning JSON

Using JSON

    string request = @"{
        ""input_file_id"": ""file-abc123"",
        ""endpoint"": ""/v1/chat/completions"",
        ""completion_window"": ""24h""
    }";
    IOpenAI openAi = new OpenAI();

    string result = await openAi.Batches.CreateJson(request);
    // result will be a JSON

Using dynamic type

    dynamic request = new {
        input_file_id = "file-abc123",
        endpoint = "/v1/chat/completions",
        completion_window = "24h"
    };
    IOpenAI openAi = new OpenAI();

    string result = await openAi.Batches.CreateJson(request);
    // result will be a JSON

Retrieve

Returning dynamic

    string batchId = "batch_abc123";
    IOpenAI openAi = new OpenAI();

    dynamic result = await openAi.Batches.RetrieveDynamic(batchId);
    // result will be a dynamic object based on the JSON result
    string id = result.id.Value;

Returning JSON

    string batchId = "batch_abc123";
    IOpenAI openAi = new OpenAI();

    string result = await openAi.Batches.RetrieveJson(batchId);
    // result will be a JSON

Cancel

Returning dynamic

    string batchId = "batch_abc123";
    IOpenAI openAi = new OpenAI();

    dynamic result = await openAi.Batches.CancelDynamic(batchId);
    // result will be a dynamic object based on the JSON result
    string id = result.id.Value;

Returning JSON

    string batchId = "batch_abc123";
    IOpenAI openAi = new OpenAI();

    string result = await openAi.Batches.CancelJson(batchId);
    // result will be a JSON

List

Returning dynamic

    IOpenAI openAi = new OpenAI();

    dynamic result = await openAi.Batches.ListDynamic();
    // result will be a dynamic object based on the JSON result
    string id = result.id.Value;

Returning JSON

    IOpenAI openAi = new OpenAI();

    string result = await openAi.Batches.ListJson();
    // result will be a JSON

Create Embeddings

Returning dynamic

Using JSON

    string request = @"{
        model: ""text-embedding-ada-002"",
        input: ""The quick brown fox jumped over the lazy dog"",
        encoding_format: ""float"",
    }";
    IOpenAI openAi = new OpenAI();

    dynamic result = await openAi.Embeddings.CreateDynamic(request);
    // result will be a dynamic array of numbers based on the JSON result
    double first = result.data[0].embedding[0].Value;

Using dynamic type

    dynamic request = new {
        model = "text-embedding-ada-002",
        input = "The quick brown fox jumped over the lazy dog",
        encoding_format = "float",
    };
    IOpenAI openAi = new OpenAI();

    dynamic result = await openAi.Embeddings.CreateDynamic(request);
    // result will be a dynamic array of numbers based on the JSON result
    double first = result.data[0].embedding[0].Value;

Returning JSON

Using JSON

    string request = @"{
        model: ""text-embedding-ada-002"",
        input: ""The quick brown fox jumped over the lazy dog"",
        encoding_format: ""float"",
    }";
    IOpenAI openAi = new OpenAI();

    string result = await openAi.Embeddings.CreateJson(request);
    // result will be a JSON

Using dynamic type

    dynamic request = new {
        model = "text-embedding-ada-002",
        input = "The quick brown fox jumped over the lazy dog",
        encoding_format = "float",
    };
    IOpenAI openAi = new OpenAI();

    string result = await openAi.Embeddings.CreateJson(request);
    // result will be a JSON

Retrieve

Returning dynamic

    string fineTuningJobId = "ftjob-abc123";
    IOpenAI openAi = new OpenAI();

    dynamic result = await openAi.FineTuning.Jobs.RetrieveDynamic(fineTuningJobId);
    // result will be a dynamic object based on the JSON result
    string id = result.id.Value;

Returning JSON

    string fineTuningJobId = "ftjob-abc123";
    IOpenAI openAi = new OpenAI();

    string result = await openAi.FineTuning.Jobs.RetrieveJson(fineTuningJobId);
    // result will be a JSON

Cancel

Returning dynamic

    string fineTuningJobId = "ftjob-abc123";
    IOpenAI openAi = new OpenAI();

    dynamic result = await openAi.FineTuning.Jobs.CancelDynamic(fineTuningJobId);
    // result will be a dynamic object based on the JSON result
    string id = result.id.Value;

Returning JSON

    string fineTuningJobId = "ftjob-abc123";
    IOpenAI openAi = new OpenAI();

    string result = await openAi.FineTuning.Jobs.CancelJson(fineTuningJobId);
    // result will be a JSON

List

Returning dynamic

    IOpenAI openAi = new OpenAI();

    dynamic result = await openAi.FineTuning.Jobs.ListDynamic();
    // result will be a dynamic object based on the JSON result
    string type = result.data[0].type.Value;

Returning JSON

    IOpenAI openAi = new OpenAI();

    string result = await openAi.FineTuning.Jobs.ListJson();
    // result will be a JSON

List Events

Returning dynamic

    string fineTuningJobId = "ftjob-abc123";
    IOpenAI openAi = new OpenAI();

    dynamic result = await openAi.FineTuning.Jobs.ListEventsDynamic(fineTuningJobId);
    // result will be a dynamic object based on the JSON result
    string type = result.data[0].type.Value;

Returning JSON

    string fineTuningJobId = "ftjob-abc123";
    IOpenAI openAi = new OpenAI();

    string result = await openAi.FineTuning.Jobs.ListEventsJson(fineTuningJobId);
    // result will be a JSON

Fine-tuning - Jobs

Create

Returning dynamic

Using JSON

    string request = @"{
      ""object"": ""fine_tuning.job"",
      ""id"": ""ftjob-abc123"",
      ""model"": ""gpt-4o-mini-2024-07-18"",
      ""created_at"": 1721764800,
      ""fine_tuned_model"": null,
      ""organization_id"": ""org-123"",
      ""result_files"": [],
      ""status"": ""queued"",
      ""validation_file"": null,
      ""training_file"": ""file-abc123"",
      ""method"": {
        ""type"": ""supervised"",
        ""supervised"": {
          ""hyperparameters"": {
            ""batch_size"": ""auto"",
            ""learning_rate_multiplier"": ""auto"",
            ""n_epochs"": ""auto"",
          }
        }
      }
    }";
    IOpenAI openAi = new OpenAI();

    dynamic result = await openAi.FineTuning.Jobs.CreateDynamic(request);
    // result will be a dynamic object based on the JSON result
    string id = result.id.Value;;

Using dynamic type

    dynamic request = new {
      object = "fine_tuning.job",
      id = "ftjob-abc123",
      model = "gpt-4o-mini-2024-07-18",
      created_at = 1721764800,
      fine_tuned_model = null,
      organization_id = "org-123",
      result_files = new [],
      status = "queued",
      validation_file = null,
      training_file = "file-abc123",
      method = new {
        type = "supervised",
        supervised = new {
          hyperparameters = new {
            batch_size = "auto",
            learning_rate_multiplier = "auto",
            n_epochs = "auto",
          }
        }
      }
    };
    IOpenAI openAi = new OpenAI();

    dynamic result = await openAi.FineTuning.Jobs.CreateDynamic(request);
    // result will be a dynamic object based on the JSON result
    string id = result.id.Value;

Returning JSON

Using JSON

    string request = @"{
      ""object"": ""fine_tuning.job"",
      ""id"": ""ftjob-abc123"",
      ""model"": ""gpt-4o-mini-2024-07-18"",
      ""created_at"": 1721764800,
      ""fine_tuned_model"": null,
      ""organization_id"": ""org-123"",
      ""result_files"": [],
      ""status"": ""queued"",
      ""validation_file"": null,
      ""training_file"": ""file-abc123"",
      ""method"": {
        ""type"": ""supervised"",
        ""supervised"": {
          ""hyperparameters"": {
            ""batch_size"": ""auto"",
            ""learning_rate_multiplier"": ""auto"",
            ""n_epochs"": ""auto"",
          }
        }
      }
    }";
    IOpenAI openAi = new OpenAI();

    string result = await openAi.FineTuning.Jobs.CreateJson(request);
    // result will be a JSON

Using dynamic type

    dynamic request = new {
      object = "fine_tuning.job",
      id = "ftjob-abc123",
      model = "gpt-4o-mini-2024-07-18",
      created_at = 1721764800,
      fine_tuned_model = null,
      organization_id = "org-123",
      result_files = new [],
      status = "queued",
      validation_file = null,
      training_file = "file-abc123",
      method = new {
        type = "supervised",
        supervised = new {
          hyperparameters = new {
            batch_size = "auto",
            learning_rate_multiplier = "auto",
            n_epochs = "auto",
          }
        }
      }
    };
    IOpenAI openAi = new OpenAI();

    string result = await openAi.FineTuning.Jobs.CreateJson(request);
    // result will be a JSON

CURL - A 'curl' command line replacement for C#.

CurlAsync

Returning a raw System.Net.Http.HttpResponseMessage

GET Method

    IOpenAI openAi = new OpenAI();
    Dictionary<string, string> headers = new Dictionary<string, string>();
    headers.Add("Authorization", "Bearer " + openAi.Settings.OpenAiApiKey);
    HttpResponseMessage response = await openAi.Curl.CurlAsync("https://api.openai.com/v1/batches/batch_abc123", "GET", headers); // "GET", null, "" or string.Empty for GET method
    string result = await response.Content.ReadAsStringAsync();
    // result will be a JSON

DELETE Method

    IOpenAI openAi = new OpenAI();
    Dictionary<string, string> headers = new Dictionary<string, string>();
    headers.Add("Authorization", "Bearer " + openAi.Settings.OpenAiApiKey);
    HttpResponseMessage response = await openAi.Curl.CurlAsync("https://api.openai.com/v1/files/file-abc123", "DELETE", headers);
    string result = await response.Content.ReadAsStringAsync();
    // result will be a JSON

POST Method

    IOpenAI openAi = new OpenAI();
    Dictionary<string, string> headers = new Dictionary<string, string>();
    headers.Add("Authorization", "Bearer " + openAi.Settings.OpenAiApiKey);
    headers.Add("Content-Type", "application/json");
    string body = @"{
        ""purpose"": ""fine-tune"",
        ""filename"": ""training_examples.jsonl"",
        ""bytes"": 2147483648,
        ""mime_type"": ""text/jsonl""
    }";
    HttpResponseMessage response = await openAi.Curl.CurlAsync("https://api.openai.com/v1/uploads", "POST", headers, body); // "POST", null, "" or string.Empty for POST method
    string result = await response.Content.ReadAsStringAsync();
    // result will be a JSON

Support this project 💖

If you find this project useful, consider supporting us by making a donation via PayPal. Your support helps keep this project alive and maintained.

• Donate to Libre-OpenAI (Beta):

Donate via QR Code

You can also scan the QR code below to make a donation: