Perplexity (#1582)

* added perplexity metadata

* added perplexity metadata

* datatype corrected

metadata/llm_metadata.yml

@@ -1,33 +1,33 @@
 openai:
   $schema: "https://json-schema.org/draft/2020-12/schema"
-  type: object
+  type: "object"
   description: "Open AI Models for Prompt"
   properties:
     temperature:
-      type: number
+      type: "number"
       default: 0.0
       minimum: 0.0
       maximum: 2.0
       description: "The temperature hyperparameter controls the creativity or randomness of the generated responses."
     max_tokens:
-      type: integer
+      type: "integer"
       default: 300
       minimum: 5
       maximum: 4096
       description: "The max_tokens hyperparameter limits the length of generated responses in chat completion using ChatGPT."
     model:
-      type: string
+      type: "string"
       default: "gpt-4o-mini"
       enum: ["gpt-3.5-turbo", "gpt-3.5-turbo-instruct", "gpt-4o-mini"]
       description: "The model hyperparameter is the ID of the model to use such as gpt-2, gpt-3, or a custom model that you have trained or fine-tuned."
     top_p:
-      type: number
+      type: "number"
       default: 0.0
       minimum: 0.0
       maximum: 1.0
       description: "The top_p hyperparameter is a value that controls the diversity of the generated responses."
     n:
-      type: integer
+      type: "integer"
       default: 1
       minimum: 1
       maximum: 5
@@ -49,67 +49,67 @@ openai:
       default: null
       description: "The stop hyperparameter is used to specify a list of tokens that should be used to indicate the end of a generated response."
     presence_penalty:
-      type: number
+      type: "number"
       default: 0.0
       minimum: -2.0
       maximum: 2.0
       description: "The presence_penalty hyperparameter penalizes the model for generating words that are not present in the context or input prompt."
     frequency_penalty:
-      type: number
+      type: "number"
       default: 0.0
       minimum: -2.0
       maximum: 2.0
       description: "The frequency_penalty hyperparameter penalizes the model for generating words that have already been generated in the current response."
     logit_bias:
-      type: object
+      type: "object"
       default: {}
       description: "The logit_bias hyperparameter helps prevent GPT-3 from generating unwanted tokens or even to encourage generation of tokens that you do want."
 anthropic:
   $schema: "https://json-schema.org/draft/2020-12/schema"
-  type: object
+  type: "object"
   description: "Anthropic AI Models for Prompt"
   properties:
     max_tokens:
-      type: integer
+      type: "integer"
       default: 1024
       minimum: 5
       maximum: 4096
       description: "The max_tokens hyperparameter limits the length of generated responses in chat completion using ChatGPT."
     model:
-      type: string
+      type: "string"
       default: "claude-3-haiku-20240307"
       enum: ["claude-3-opus-20240229", "claude-3-5-sonnet-20240620", "claude-3-sonnet-20240229", "claude-3-haiku-20240307"]
       description: "The model hyperparameter is the ID of the Anthropic model or a custom model that you have trained or fine-tuned."
 gemini:
   $schema: "https://json-schema.org/draft/2020-12/schema"
-  type: object
+  type: "object"
   description: "Google AI Models for Prompt"
   properties:
     temperature:
-      type: number
+      type: "number"
       default: 0.0
       minimum: 0.0
       maximum: 2.0
       description: "The temperature hyperparameter controls the creativity or randomness of the generated responses."
     max_tokens:
-      type: integer
+      type: "integer"
       default: 300
       minimum: 5
       maximum: 4096
       description: "The max_tokens hyperparameter limits the length of generated responses in chat completion."
     model:
-      type: string
+      type: "string"
       default: "gemini/gemini-1.5-flash"
       enum: ["gemini/gemini-1.5-flash", "gemini/gemini-pro"]
       description: "The model hyperparameter is the ID of the model to use."
     top_p:
-      type: number
+      type: "number"
       default: 0.0
       minimum: 0.0
       maximum: 1.0
       description: "The top_p hyperparameter is a value that controls the diversity of the generated responses."
     n:
-      type: integer
+      type: "integer"
       default: 1
       minimum: 1
       maximum: 5
@@ -129,4 +129,76 @@ gemini:
         - "integer"
         - "null"
       default: null
-      description: "The stop hyperparameter is used to specify a list of tokens that should be used to indicate the end of a generated response."
+      description: "The stop hyperparameter is used to specify a list of tokens that should be used to indicate the end of a generated response."
+perplexity:
+  $schema: "https://json-schema.org/draft/2020-12/schema"
+  type: "object"
+  description: "Perplexity AI Models for Prompt"
+  properties:
+    temperature:
+      type: "number"
+      default: 0.2
+      minimum: 0.0
+      maximum: 2.0
+      description: "The temperature hyperparameter controls the creativity or randomness of the generated responses."
+    max_tokens:
+      anyOf:
+        - type: "integer"
+          minimum: 5
+          maximum: 127072
+        - type: "null"
+      type:
+        - "integer"
+        - "null"
+      default: null
+      description: "The max_tokens hyperparameter limits the length of generated responses in chat completion"
+    model:
+      type: "string"
+      default: "perplexity/llama-3.1-sonar-small-128k-online"
+      enum: ["perplexity/llama-3.1-sonar-small-128k-online", "perplexity/llama-3.1-sonar-large-128k-online", "perplexity/llama-3.1-sonar-huge-128k-online"]
+    search_domain_filter:
+      anyOf:
+        - type: "array"
+          maxItems: 3
+          items:
+            type: "string"
+        - type: "null"
+      type:
+        - "array"
+        - "null"
+      default: null
+      description: "The search domain filter hyperparameter is used to specify list of domain to be used by online models."
+    search_recency_filter:
+      anyOf:
+        - type: "string"
+          enum: ["month", "week", "day", "hour"]
+        - type: "null"
+      type:
+        - "string"
+        - "null"
+      default: null
+      description: "return results from specified time interval"
+    top_p:
+      type: "number"
+      default: 0.9
+      minimum: 0.0
+      maximum: 1.0
+      description: "The top_p hyperparameter is a value that controls the diversity of the generated responses."
+    top_k:
+      type: "integer"
+      default: 0
+      minimum: 0
+      maximum: 2048
+      description: "The top_k hyperparameter controls the number of token to keep for top_k filtering"
+    presence_penalty:
+      type: "number"
+      default: 0.0
+      minimum: -2.0
+      maximum: 2.0
+      description: "The presence_penalty hyperparameter penalizes the model for generating words that are not present in the context or input prompt."
+    frequency_penalty:
+      type: "number"
+      default: 0.0
+      minimum: -2.0
+      maximum: 2.0
+      description: "The frequency_penalty hyperparameter penalizes the model for generating words that have already been generated in the current response."