chore: engine model adding ux refinement (#4580)
* fix: enhance remote engine models adding UX * fix: OpenAI o3 support template * chore: fix models populating error * chore: missing dependency
This commit is contained in:
Louis
GitHub
parent
095a00441c
commit
4846befd4a
@ -18,6 +18,7 @@ export type EngineMetadata = {
|
||||
template?: string
|
||||
}
|
||||
}
|
||||
explore_models_url?: string
|
||||
}
|
||||
|
||||
export type EngineVariant = {
|
||||
|
||||
@ -17,6 +17,7 @@
|
||||
"chat_completions": {
|
||||
"template": "{% if input_request.stream %} {\"object\": \"chat.completion.chunk\", \"model\": \"{{ input_request.model }}\", \"choices\": [{\"index\": 0, \"delta\": { {% if input_request.type == \"message_start\" %} \"role\": \"assistant\", \"content\": null {% else if input_request.type == \"ping\" %} \"role\": \"assistant\", \"content\": null {% else if input_request.type == \"content_block_delta\" %} \"role\": \"assistant\", \"content\": \"{{ input_request.delta.text }}\" {% else if input_request.type == \"content_block_stop\" %} \"role\": \"assistant\", \"content\": null {% else if input_request.type == \"content_block_stop\" %} \"role\": \"assistant\", \"content\": null {% endif %} }, {% if input_request.type == \"content_block_stop\" %} \"finish_reason\": \"stop\" {% else %} \"finish_reason\": null {% endif %} }]} {% else %} {\"id\": \"{{ input_request.id }}\", \"created\": null, \"object\": \"chat.completion\", \"model\": \"{{ input_request.model }}\", \"choices\": [{ \"index\": 0, \"message\": { \"role\": \"{{ input_request.role }}\", \"content\": {% if input_request.content and input_request.content.0.type == \"text\" %} \"{{input_request.content.0.text}}\" {% else %} null {% endif %}, \"refusal\": null }, \"logprobs\": null, \"finish_reason\": \"{{ input_request.stop_reason }}\" } ], \"usage\": { \"prompt_tokens\": {{ input_request.usage.input_tokens }}, \"completion_tokens\": {{ input_request.usage.output_tokens }}, \"total_tokens\": {{ input_request.usage.input_tokens + input_request.usage.output_tokens }}, \"prompt_tokens_details\": { \"cached_tokens\": 0 }, \"completion_tokens_details\": { \"reasoning_tokens\": 0, \"accepted_prediction_tokens\": 0, \"rejected_prediction_tokens\": 0 } }, \"system_fingerprint\": \"fp_6b68a8204b\"} {% endif %}"
|
||||
}
|
||||
}
|
||||
},
|
||||
"explore_models_url": "https://docs.anthropic.com/en/docs/about-claude/models"
|
||||
}
|
||||
}
|
||||
|
||||
@ -17,6 +17,7 @@
|
||||
"chat_completions": {
|
||||
"template": "{% if input_request.stream %} {\"object\": \"chat.completion.chunk\", \"model\": \"{{ input_request.model }}\", \"choices\": [{\"index\": 0, \"delta\": { {% if input_request.event_type == \"text-generation\" %} \"role\": \"assistant\", \"content\": \"{{ input_request.text }}\" {% else %} \"role\": \"assistant\", \"content\": null {% endif %} }, {% if input_request.event_type == \"stream-end\" %} \"finish_reason\": \"{{ input_request.finish_reason }}\" {% else %} \"finish_reason\": null {% endif %} }]} {% else %} {\"id\": \"{{ input_request.generation_id }}\", \"created\": null, \"object\": \"chat.completion\", \"model\": {% if input_request.model %} \"{{ input_request.model }}\" {% else %} \"command-r-plus-08-2024\" {% endif %}, \"choices\": [{ \"index\": 0, \"message\": { \"role\": \"assistant\", \"content\": {% if not input_request.text %} null {% else %} \"{{ input_request.text }}\" {% endif %}, \"refusal\": null }, \"logprobs\": null, \"finish_reason\": \"{{ input_request.finish_reason }}\" } ], \"usage\": { \"prompt_tokens\": {{ input_request.meta.tokens.input_tokens }}, \"completion_tokens\": {{ input_request.meta.tokens.output_tokens }},\"total_tokens\": {{ input_request.meta.tokens.input_tokens + input_request.meta.tokens.output_tokens }}, \"prompt_tokens_details\": { \"cached_tokens\": 0 },\"completion_tokens_details\": { \"reasoning_tokens\": 0, \"accepted_prediction_tokens\": 0, \"rejected_prediction_tokens\": 0 } }, \"system_fingerprint\": \"fp_6b68a8204b\"} {% endif %}"
|
||||
}
|
||||
}
|
||||
},
|
||||
"explore_models_url": "https://docs.cohere.com/v2/docs/models"
|
||||
}
|
||||
}
|
||||
|
||||
@ -17,6 +17,7 @@
|
||||
"chat_completions": {
|
||||
"template": "{ {% set first = true %} {% for key, value in input_request %} {% if key == \"choices\" or key == \"created\" or key == \"model\" or key == \"service_tier\" or key == \"system_fingerprint\" or key == \"stream\" or key == \"object\" or key == \"usage\" %} {% if not first %},{% endif %} \"{{ key }}\": {{ tojson(value) }} {% set first = false %} {% endif %} {% endfor %} }"
|
||||
}
|
||||
}
|
||||
},
|
||||
"explore_models_url": "https://console.groq.com/docs/models"
|
||||
}
|
||||
}
|
||||
|
||||
@ -17,6 +17,7 @@
|
||||
"chat_completions": {
|
||||
"template": "{ {% set first = true %} {% for key, value in input_request %} {% if key == \"choices\" or key == \"created\" or key == \"model\" or key == \"service_tier\" or key == \"system_fingerprint\" or key == \"stream\" or key == \"object\" or key == \"usage\" %} {% if not first %},{% endif %} \"{{ key }}\": {{ tojson(value) }} {% set first = false %} {% endif %} {% endfor %} }"
|
||||
}
|
||||
}
|
||||
},
|
||||
"explore_models_url": "https://withmartian.github.io/llm-adapters/"
|
||||
}
|
||||
}
|
||||
|
||||
@ -17,6 +17,7 @@
|
||||
"chat_completions": {
|
||||
"template": "{{tojson(input_request)}}"
|
||||
}
|
||||
}
|
||||
},
|
||||
"explore_models_url": "https://docs.mistral.ai/getting-started/models/models_overview/"
|
||||
}
|
||||
}
|
||||
|
||||
@ -17,6 +17,7 @@
|
||||
"chat_completions": {
|
||||
"template": "{ {% set first = true %} {% for key, value in input_request %} {% if key == \"choices\" or key == \"created\" or key == \"model\" or key == \"service_tier\" or key == \"system_fingerprint\" or key == \"stream\" or key == \"object\" or key == \"usage\" %} {% if not first %},{% endif %} \"{{ key }}\": {{ tojson(value) }} {% set first = false %} {% endif %} {% endfor %} }"
|
||||
}
|
||||
}
|
||||
},
|
||||
"explore_models_url": "https://build.nvidia.com/models"
|
||||
}
|
||||
}
|
||||
|
||||
@ -10,13 +10,14 @@
|
||||
"transform_req": {
|
||||
"chat_completions": {
|
||||
"url": "https://api.openai.com/v1/chat/completions",
|
||||
"template": "{ {% set first = true %} {% for key, value in input_request %} {% if key == \"model\" or key == \"temperature\" or key == \"store\" or key == \"messages\" or key == \"stream\" or key == \"presence_penalty\" or key == \"metadata\" or key == \"frequency_penalty\" or key == \"tools\" or key == \"tool_choice\" or key == \"logprobs\" or key == \"top_logprobs\" or key == \"logit_bias\" or key == \"n\" or key == \"modalities\" or key == \"prediction\" or key == \"response_format\" or key == \"service_tier\" or key == \"seed\" or key == \"stream_options\" or key == \"top_p\" or key == \"parallel_tool_calls\" or key == \"user\" or key == \"max_tokens\" or ((input_request.model == \"o1\" or input_request.model == \"o1-preview\" or input_request.model == \"o1-mini\") and (key == \"stop\")) %} {% if not first %} , {% endif %} {% if key == \"messages\" and (input_request.model == \"o1\" or input_request.model == \"o1-preview\" or input_request.model == \"o1-mini\") and input_request.messages.0.role == \"system\" %} \"messages\": [{% for message in input_request.messages %} {% if not loop.is_first %} { \"role\": \"{{ message.role }}\", \"content\": \"{{ message.content }}\" } {% if not loop.is_last %} , {% endif %} {% endif %} {% endfor %}] {% else if key == \"max_tokens\" and (input_request.model == \"o1\" or input_request.model == \"o1-preview\" or input_request.model == \"o1-mini\") %} \"max_completion_tokens\": {{ tojson(value) }} {% else %} \"{{ key }}\": {{ tojson(value) }} {% set first = false %} {% endif %} {% endif %} {% endfor %} }"
|
||||
"template": "{ {% set first = true %} {% for key, value in input_request %} {% if key == \"model\" or key == \"temperature\" or key == \"store\" or key == \"messages\" or key == \"stream\" or key == \"presence_penalty\" or key == \"metadata\" or key == \"frequency_penalty\" or key == \"tools\" or key == \"tool_choice\" or key == \"logprobs\" or key == \"top_logprobs\" or key == \"logit_bias\" or key == \"n\" or key == \"modalities\" or key == \"prediction\" or key == \"response_format\" or key == \"service_tier\" or key == \"seed\" or key == \"stream_options\" or key == \"top_p\" or key == \"parallel_tool_calls\" or key == \"user\" or key == \"max_tokens\" or ((input_request.model == \"o1\" or input_request.model == \"o1-preview\" or input_request.model == \"o1-mini\" or input_request.model == \"o3\" or input_request.model == \"o3-mini\") and (key == \"stop\")) %} {% if not first %} , {% endif %} {% if key == \"messages\" and (input_request.model == \"o1\" or input_request.model == \"o1-preview\" or input_request.model == \"o1-mini\") and input_request.messages.0.role == \"system\" %} \"messages\": [{% for message in input_request.messages %} {% if not loop.is_first %} { \"role\": \"{{ message.role }}\", \"content\": \"{{ message.content }}\" } {% if not loop.is_last %} , {% endif %} {% endif %} {% endfor %}] {% else if key == \"max_tokens\" and (input_request.model == \"o1\" or input_request.model == \"o1-preview\" or input_request.model == \"o1-mini\" or input_request.model == \"o3\" or input_request.model == \"o3-mini\") %} \"max_completion_tokens\": {{ tojson(value) }} {% else %} \"{{ key }}\": {{ tojson(value) }} {% set first = false %} {% endif %} {% endif %} {% endfor %} }"
|
||||
}
|
||||
},
|
||||
"transform_resp": {
|
||||
"chat_completions": {
|
||||
"template": "{ {% set first = true %} {% for key, value in input_request %} {% if key == \"choices\" or key == \"created\" or key == \"model\" or key == \"service_tier\" or key == \"stream\" or key == \"object\" or key == \"usage\" %} {% if not first %},{% endif %} \"{{ key }}\": {{ tojson(value) }} {% set first = false %} {% endif %} {% endfor %} }"
|
||||
}
|
||||
}
|
||||
},
|
||||
"explore_models_url": "https://platform.openai.com/docs/models"
|
||||
}
|
||||
}
|
||||
|
||||
@ -17,6 +17,7 @@
|
||||
"chat_completions": {
|
||||
"template": "{ {% set first = true %} {% for key, value in input_request %} {% if key == \"choices\" or key == \"created\" or key == \"model\" or key == \"service_tier\" or key == \"system_fingerprint\" or key == \"stream\" or key == \"object\" or key == \"usage\" %} {% if not first %},{% endif %} \"{{ key }}\": {{ tojson(value) }} {% set first = false %} {% endif %} {% endfor %} }"
|
||||
}
|
||||
}
|
||||
},
|
||||
"explore_models_url": "https://openrouter.ai/models"
|
||||
}
|
||||
}
|
||||
|
||||
@ -167,7 +167,9 @@ export default class JSONEngineManagementExtension extends EngineManagementExten
|
||||
ky.post(`${API_URL}/v1/engines`, { json: engineConfig }).then((e) => {
|
||||
if (persistModels && engineConfig.metadata?.get_models_url) {
|
||||
// Pull /models from remote models endpoint
|
||||
return this.populateRemoteModels(engineConfig).then(() => e)
|
||||
return this.populateRemoteModels(engineConfig)
|
||||
.then(() => e)
|
||||
.catch(() => e)
|
||||
}
|
||||
return e
|
||||
})
|
||||
|
||||
@ -293,6 +293,7 @@ const ModelDropdown = ({
|
||||
updateThreadMetadata,
|
||||
setThreadModelParams,
|
||||
updateModelParameter,
|
||||
stopModel,
|
||||
]
|
||||
)
|
||||
|
||||
|
||||
@ -30,7 +30,7 @@
|
||||
"jotai": "^2.6.0",
|
||||
"katex": "^0.16.10",
|
||||
"lodash": "^4.17.21",
|
||||
"lucide-react": "^0.291.0",
|
||||
"lucide-react": "^0.311.0",
|
||||
"marked": "^9.1.2",
|
||||
"next": "14.2.3",
|
||||
"next-themes": "^0.2.1",
|
||||
|
||||
@ -10,7 +10,7 @@ import { InferenceEngine, Model } from '@janhq/core'
|
||||
|
||||
import { Button, Input, Modal } from '@janhq/joi'
|
||||
import { useAtomValue } from 'jotai'
|
||||
import { PlusIcon } from 'lucide-react'
|
||||
import { PlusIcon, ArrowUpRightFromSquare } from 'lucide-react'
|
||||
|
||||
import { z } from 'zod'
|
||||
|
||||
@ -71,7 +71,7 @@ const ModelAddModel = ({ engine }: { engine: string }) => {
|
||||
{prefix}
|
||||
{label}
|
||||
</span>
|
||||
<p className="mt-1 font-normal text-[hsla(var(--text-secondary))]">
|
||||
<p className="mt-4 font-normal text-[hsla(var(--text-secondary))]">
|
||||
{desc}
|
||||
{isRequired && <span className="text-red-500">*</span>}
|
||||
</p>
|
||||
@ -97,7 +97,7 @@ const ModelAddModel = ({ engine }: { engine: string }) => {
|
||||
className="w-[500px]"
|
||||
content={
|
||||
<div>
|
||||
<form className="mt-8 space-y-6" onSubmit={handleSubmit(onSubmit)}>
|
||||
<form className="mt-4 space-y-6" onSubmit={handleSubmit(onSubmit)}>
|
||||
<div className="space-y-2">
|
||||
<label htmlFor="modelName" className="font-semibold">
|
||||
{renderLabel(
|
||||
@ -123,14 +123,19 @@ const ModelAddModel = ({ engine }: { engine: string }) => {
|
||||
{errors.modelName.message}
|
||||
</p>
|
||||
)}
|
||||
<div className="pt-4">
|
||||
<div className="pt-2">
|
||||
<a
|
||||
target="_blank"
|
||||
href={engines?.[engine as InferenceEngine]?.[0]?.url}
|
||||
className="text-[hsla(var(--app-link))]"
|
||||
href={
|
||||
engines?.[engine as InferenceEngine]?.[0]?.metadata
|
||||
?.explore_models_url ??
|
||||
engines?.[engine as InferenceEngine]?.[0]?.url
|
||||
}
|
||||
className="flex flex-row items-center gap-1 font-medium text-[hsla(var(--app-link))] no-underline"
|
||||
>
|
||||
See model list from{' '}
|
||||
{getTitleByEngine(engine as InferenceEngine)}
|
||||
<ArrowUpRightFromSquare size={13} />
|
||||
</a>
|
||||
</div>
|
||||
</div>
|
||||
|
||||
Loading…
x
Reference in New Issue
Block a user