📊 refactor: use Parameters from Side Panel for OpenAI, Anthropic, and Custom endpoints (#4092)

* feat: openai parameters

* refactor: anthropic/bedrock params, add preset params for openai, and add azure params

* refactor: use 'compact' schemas for anthropic/openai

* refactor: ensure custom endpoints are properly recognized as valid param endpoints

* refactor: update paramEndpoints check in BaseClient.js

* chore: optimize logging by omitting modelsConfig

* refactor: update label casing in baseDefinitions combobox items

* fix: remove 'stop' model options when using o1 series models

* refactor(AnthropicClient): remove default `stop` value

* refactor: reset params on parameters change

* refactor: remove unused default parameter value map introduced in prior commit

* fix: 'min' typo for 'max' value

* refactor: preset settings

* refactor: replace dropdown for image detail with slider; remove `preventDelayedUpdate` condition from DynamicSlider

* fix: localizations for freq./pres. penalty

* Refactor maxOutputTokens to use coerceNumber in tConversationSchema

* refactor(AnthropicClient): use `getModelMaxOutputTokens`

api/utils/tokens.js

@@ -123,7 +123,16 @@ const modelMaxOutputs = {
   system_default: 1024,
 };
 
+const anthropicMaxOutputs = {
+  'claude-3-haiku': 4096,
+  'claude-3-sonnet': 4096,
+  'claude-3-opus': 4096,
+  'claude-3.5-sonnet': 8192,
+  'claude-3-5-sonnet': 8192,
+};
+
 const maxOutputTokensMap = {
+  [EModelEndpoint.anthropic]: anthropicMaxOutputs,
   [EModelEndpoint.azureOpenAI]: modelMaxOutputs,
   [EModelEndpoint.openAI]: modelMaxOutputs,
   [EModelEndpoint.custom]: modelMaxOutputs,