-
-
- {t("modelSettings.label")}
-
+
+
+ {t("modelSettings.label")}
+
{t("modelSettings.description")}
@@ -41,6 +40,7 @@ export const ModelSettings = () => {
temperature: number
topK: number
topP: number
+ numGpu: number
}) => {
Object.entries(values).forEach(([key, value]) => {
setModelSetting(key, value)
@@ -106,6 +106,17 @@ export const ModelSettings = () => {
placeholder={t("modelSettings.form.topP.placeholder")}
/>
+
+
+
)
}
diff --git a/src/components/Option/Settings/rag.tsx b/src/components/Option/Settings/rag.tsx
index 12d7d97..fd5d576 100644
--- a/src/components/Option/Settings/rag.tsx
+++ b/src/components/Option/Settings/rag.tsx
@@ -10,7 +10,7 @@ import {
} from "~/services/ollama"
import { SettingPrompt } from "./prompt"
import { useTranslation } from "react-i18next"
-import { getTotalFilePerKB } from "@/services/app"
+import { getNoOfRetrievedDocs, getTotalFilePerKB } from "@/services/app"
import { SidepanelRag } from "./sidepanel-rag"
export const RagSettings = () => {
@@ -21,20 +21,28 @@ export const RagSettings = () => {
const { data: ollamaInfo, status } = useQuery({
queryKey: ["fetchRAGSettings"],
queryFn: async () => {
- const [allModels, chunkOverlap, chunkSize, defaultEM, totalFilePerKB] =
- await Promise.all([
- getAllModels({ returnEmpty: true }),
- defaultEmbeddingChunkOverlap(),
- defaultEmbeddingChunkSize(),
- defaultEmbeddingModelForRag(),
- getTotalFilePerKB()
- ])
+ const [
+ allModels,
+ chunkOverlap,
+ chunkSize,
+ defaultEM,
+ totalFilePerKB,
+ noOfRetrievedDocs
+ ] = await Promise.all([
+ getAllModels({ returnEmpty: true }),
+ defaultEmbeddingChunkOverlap(),
+ defaultEmbeddingChunkSize(),
+ defaultEmbeddingModelForRag(),
+ getTotalFilePerKB(),
+ getNoOfRetrievedDocs()
+ ])
return {
models: allModels,
chunkOverlap,
chunkSize,
defaultEM,
- totalFilePerKB
+ totalFilePerKB,
+ noOfRetrievedDocs
}
}
})
@@ -45,8 +53,15 @@ export const RagSettings = () => {
chunkSize: number
overlap: number
totalFilePerKB: number
+ noOfRetrievedDocs: number
}) => {
- await saveForRag(data.model, data.chunkSize, data.overlap, data.totalFilePerKB)
+ await saveForRag(
+ data.model,
+ data.chunkSize,
+ data.overlap,
+ data.totalFilePerKB,
+ data.noOfRetrievedDocs
+ )
return true
},
onSuccess: () => {
@@ -75,14 +90,16 @@ export const RagSettings = () => {
model: data.defaultEM,
chunkSize: data.chunkSize,
overlap: data.chunkOverlap,
- totalFilePerKB: data.totalFilePerKB
+ totalFilePerKB: data.totalFilePerKB,
+ noOfRetrievedDocs: data.noOfRetrievedDocs
})
}}
initialValues={{
chunkSize: ollamaInfo?.chunkSize,
chunkOverlap: ollamaInfo?.chunkOverlap,
defaultEM: ollamaInfo?.defaultEM,
- totalFilePerKB: ollamaInfo?.totalFilePerKB
+ totalFilePerKB: ollamaInfo?.totalFilePerKB,
+ noOfRetrievedDocs: ollamaInfo?.noOfRetrievedDocs
}}>
{
/>
+
+
+
+
{
]}>
-
diff --git a/src/entries/background.ts b/src/entries/background.ts
index 1f0de4b..9a9bcd5 100644
--- a/src/entries/background.ts
+++ b/src/entries/background.ts
@@ -64,31 +64,31 @@ export default defineBackground({
browser.contextMenus.create({
id: "summarize-pa",
- title: "Summarize",
+ title: browser.i18n.getMessage("contextSummarize"),
contexts: ["selection"]
})
browser.contextMenus.create({
id: "explain-pa",
- title: "Explain",
+ title: browser.i18n.getMessage("contextExplain"),
contexts: ["selection"]
})
browser.contextMenus.create({
id: "rephrase-pa",
- title: "Rephrase",
+ title: browser.i18n.getMessage("contextRephrase"),
contexts: ["selection"]
})
browser.contextMenus.create({
id: "translate-pg",
- title: "Translate",
+ title: browser.i18n.getMessage("contextTranslate"),
contexts: ["selection"]
})
browser.contextMenus.create({
id: "custom-pg",
- title: "Custom",
+ title: browser.i18n.getMessage("contextCustom"),
contexts: ["selection"]
})
diff --git a/src/entries/options/index.html b/src/entries/options/index.html
index 05423c3..c30dfad 100644
--- a/src/entries/options/index.html
+++ b/src/entries/options/index.html
@@ -4,6 +4,7 @@
Page Assist - A Web UI for Local AI Models
+
diff --git a/src/hooks/useMessage.tsx b/src/hooks/useMessage.tsx
index 3d61528..abf9a6b 100644
--- a/src/hooks/useMessage.tsx
+++ b/src/hooks/useMessage.tsx
@@ -122,7 +122,9 @@ export const useMessage = () => {
topP: currentChatModelSettings?.topP ?? userDefaultModelSettings?.topP,
numCtx:
currentChatModelSettings?.numCtx ?? userDefaultModelSettings?.numCtx,
- seed: currentChatModelSettings?.seed
+ seed: currentChatModelSettings?.seed,
+ numGpu:
+ currentChatModelSettings?.numGpu ?? userDefaultModelSettings?.numGpu
})
let newMessage: Message[] = []
@@ -248,7 +250,9 @@ export const useMessage = () => {
numCtx:
currentChatModelSettings?.numCtx ??
userDefaultModelSettings?.numCtx,
- seed: currentChatModelSettings?.seed
+ seed: currentChatModelSettings?.seed,
+ numGpu:
+ currentChatModelSettings?.numGpu ?? userDefaultModelSettings?.numGpu
})
const response = await questionOllama.invoke(promptForQuestion)
query = response.content.toString()
@@ -442,7 +446,9 @@ export const useMessage = () => {
topP: currentChatModelSettings?.topP ?? userDefaultModelSettings?.topP,
numCtx:
currentChatModelSettings?.numCtx ?? userDefaultModelSettings?.numCtx,
- seed: currentChatModelSettings?.seed
+ seed: currentChatModelSettings?.seed,
+ numGpu:
+ currentChatModelSettings?.numGpu ?? userDefaultModelSettings?.numGpu
})
let newMessage: Message[] = []
@@ -645,7 +651,9 @@ export const useMessage = () => {
topP: currentChatModelSettings?.topP ?? userDefaultModelSettings?.topP,
numCtx:
currentChatModelSettings?.numCtx ?? userDefaultModelSettings?.numCtx,
- seed: currentChatModelSettings?.seed
+ seed: currentChatModelSettings?.seed,
+ numGpu:
+ currentChatModelSettings?.numGpu ?? userDefaultModelSettings?.numGpu
})
let newMessage: Message[] = []
@@ -718,7 +726,9 @@ export const useMessage = () => {
numCtx:
currentChatModelSettings?.numCtx ??
userDefaultModelSettings?.numCtx,
- seed: currentChatModelSettings?.seed
+ seed: currentChatModelSettings?.seed,
+ numGpu:
+ currentChatModelSettings?.numGpu ?? userDefaultModelSettings?.numGpu
})
const response = await questionOllama.invoke(promptForQuestion)
query = response.content.toString()
@@ -890,7 +900,9 @@ export const useMessage = () => {
topP: currentChatModelSettings?.topP ?? userDefaultModelSettings?.topP,
numCtx:
currentChatModelSettings?.numCtx ?? userDefaultModelSettings?.numCtx,
- seed: currentChatModelSettings?.seed
+ seed: currentChatModelSettings?.seed,
+ numGpu:
+ currentChatModelSettings?.numGpu ?? userDefaultModelSettings?.numGpu
})
let newMessage: Message[] = []
@@ -932,8 +944,6 @@ export const useMessage = () => {
let contentToSave = ""
try {
-
-
const prompt = await getPrompt(messageType)
let humanMessage = new HumanMessage({
content: [
diff --git a/src/hooks/useMessageOption.tsx b/src/hooks/useMessageOption.tsx
index 2b28aa7..7dd6d6d 100644
--- a/src/hooks/useMessageOption.tsx
+++ b/src/hooks/useMessageOption.tsx
@@ -32,6 +32,7 @@ import { useStorage } from "@plasmohq/storage/hook"
import { useStoreChatModelSettings } from "@/store/model"
import { getAllDefaultModelSettings } from "@/services/model-settings"
import { pageAssistModel } from "@/models"
+import { getNoOfRetrievedDocs } from "@/services/app"
export const useMessageOption = () => {
const {
@@ -117,7 +118,9 @@ export const useMessageOption = () => {
topP: currentChatModelSettings?.topP ?? userDefaultModelSettings?.topP,
numCtx:
currentChatModelSettings?.numCtx ?? userDefaultModelSettings?.numCtx,
- seed: currentChatModelSettings?.seed
+ seed: currentChatModelSettings?.seed,
+ numGpu:
+ currentChatModelSettings?.numGpu ?? userDefaultModelSettings?.numGpu
})
let newMessage: Message[] = []
@@ -190,7 +193,9 @@ export const useMessageOption = () => {
numCtx:
currentChatModelSettings?.numCtx ??
userDefaultModelSettings?.numCtx,
- seed: currentChatModelSettings?.seed
+ seed: currentChatModelSettings?.seed,
+ numGpu:
+ currentChatModelSettings?.numGpu ?? userDefaultModelSettings?.numGpu
})
const response = await questionOllama.invoke(promptForQuestion)
query = response.content.toString()
@@ -360,7 +365,9 @@ export const useMessageOption = () => {
topP: currentChatModelSettings?.topP ?? userDefaultModelSettings?.topP,
numCtx:
currentChatModelSettings?.numCtx ?? userDefaultModelSettings?.numCtx,
- seed: currentChatModelSettings?.seed
+ seed: currentChatModelSettings?.seed,
+ numGpu:
+ currentChatModelSettings?.numGpu ?? userDefaultModelSettings?.numGpu
})
let newMessage: Message[] = []
@@ -576,7 +583,9 @@ export const useMessageOption = () => {
topP: currentChatModelSettings?.topP ?? userDefaultModelSettings?.topP,
numCtx:
currentChatModelSettings?.numCtx ?? userDefaultModelSettings?.numCtx,
- seed: currentChatModelSettings?.seed
+ seed: currentChatModelSettings?.seed,
+ numGpu:
+ currentChatModelSettings?.numGpu ?? userDefaultModelSettings?.numGpu
})
let newMessage: Message[] = []
@@ -665,13 +674,16 @@ export const useMessageOption = () => {
numCtx:
currentChatModelSettings?.numCtx ??
userDefaultModelSettings?.numCtx,
- seed: currentChatModelSettings?.seed
+ seed: currentChatModelSettings?.seed,
+ numGpu:
+ currentChatModelSettings?.numGpu ?? userDefaultModelSettings?.numGpu
})
const response = await questionOllama.invoke(promptForQuestion)
query = response.content.toString()
}
+ const docSize = await getNoOfRetrievedDocs()
- const docs = await vectorstore.similaritySearch(query, 4)
+ const docs = await vectorstore.similaritySearch(query, docSize)
const context = formatDocs(docs)
const source = docs.map((doc) => {
return {
diff --git a/src/models/ChatOllama.ts b/src/models/ChatOllama.ts
index 2a4c28b..1bc7ddb 100644
--- a/src/models/ChatOllama.ts
+++ b/src/models/ChatOllama.ts
@@ -126,7 +126,7 @@ export class ChatOllama
this.mirostatTau = fields.mirostatTau;
this.numBatch = fields.numBatch;
this.numCtx = fields.numCtx;
- this.numGpu = fields.numGpu;
+ this.numGpu = fields.numGpu === null ? undefined : fields.numGpu;
this.numGqa = fields.numGqa;
this.numKeep = fields.numKeep;
this.numPredict = fields.numPredict;
diff --git a/src/models/index.ts b/src/models/index.ts
index 4048cb3..ce3ab39 100644
--- a/src/models/index.ts
+++ b/src/models/index.ts
@@ -9,7 +9,8 @@ export const pageAssistModel = async ({
topK,
topP,
numCtx,
- seed
+ seed,
+ numGpu
}: {
model: string
baseUrl: string
@@ -19,6 +20,7 @@ export const pageAssistModel = async ({
topP?: number
numCtx?: number
seed?: number
+ numGpu?: number
}) => {
switch (model) {
case "chrome::gemini-nano::page-assist":
@@ -35,7 +37,8 @@ export const pageAssistModel = async ({
topP,
numCtx,
seed,
- model
+ model,
+ numGpu
})
}
}
diff --git a/src/public/_locales/en/messages.json b/src/public/_locales/en/messages.json
index a574d20..cd34708 100644
--- a/src/public/_locales/en/messages.json
+++ b/src/public/_locales/en/messages.json
@@ -10,5 +10,20 @@
},
"openOptionToChat": {
"message": "Open Web UI to Chat"
+ },
+ "contextSummarize": {
+ "message": "Summarize"
+ },
+ "contextExplain": {
+ "message": "Explain"
+ },
+ "contextRephrase": {
+ "message": "Rephrase"
+ },
+ "contextTranslate" :{
+ "message": "Translate"
+ },
+ "contextCustom": {
+ "message": "Custom"
}
}
\ No newline at end of file
diff --git a/src/public/_locales/es/messages.json b/src/public/_locales/es/messages.json
index a7272c7..af086ab 100644
--- a/src/public/_locales/es/messages.json
+++ b/src/public/_locales/es/messages.json
@@ -7,5 +7,23 @@
},
"openSidePanelToChat": {
"message": "Abrir Copilot para Chatear"
+ },
+ "openOptionToChat": {
+ "message": "Abrir Web UI para Chatear"
+ },
+ "contextSummarize": {
+ "message": "Resumir"
+ },
+ "contextExplain": {
+ "message": "Explicar"
+ },
+ "contextRephrase": {
+ "message": "Reformular"
+ },
+ "contextTranslate": {
+ "message": "Traducir"
+ },
+ "contextCustom": {
+ "message": "Personalizado"
}
-}
+}
\ No newline at end of file
diff --git a/src/public/_locales/fa/messages.json b/src/public/_locales/fa/messages.json
index 2d047ed..c193a32 100644
--- a/src/public/_locales/fa/messages.json
+++ b/src/public/_locales/fa/messages.json
@@ -10,5 +10,20 @@
},
"openOptionToChat": {
"message": "باز کردن رابط کاربری وب برای گفتگو"
+ },
+ "contextSummarize": {
+ "message": "خلاصه کردن"
+ },
+ "contextExplain": {
+ "message": "توضیح دادن"
+ },
+ "contextRephrase": {
+ "message": "بازنویسی"
+ },
+ "contextTranslate" :{
+ "message": "ترجمه کردن"
+ },
+ "contextCustom": {
+ "message": "سفارشی"
}
}
diff --git a/src/public/_locales/fr/messages.json b/src/public/_locales/fr/messages.json
index 6b53932..179752d 100644
--- a/src/public/_locales/fr/messages.json
+++ b/src/public/_locales/fr/messages.json
@@ -7,5 +7,20 @@
},
"openSidePanelToChat": {
"message": "Ouvrir Copilot pour discuter"
+ },
+ "contextSummarize": {
+ "message": "Résumer"
+ },
+ "contextExplain": {
+ "message": "Expliquer"
+ },
+ "contextRephrase": {
+ "message": "Reformuler"
+ },
+ "contextTranslate" :{
+ "message": "Traduire"
+ },
+ "contextCustom": {
+ "message": "Personnalisé"
}
}
\ No newline at end of file
diff --git a/src/public/_locales/it/messages.json b/src/public/_locales/it/messages.json
index eb773d3..eaee088 100644
--- a/src/public/_locales/it/messages.json
+++ b/src/public/_locales/it/messages.json
@@ -7,5 +7,20 @@
},
"openSidePanelToChat": {
"message": "Apri Copilot per chattare"
+ },
+ "contextSummarize": {
+ "message": "Riassumi"
+ },
+ "contextExplain": {
+ "message": "Spiega"
+ },
+ "contextRephrase": {
+ "message": "Riformula"
+ },
+ "contextTranslate" :{
+ "message": "Traduci"
+ },
+ "contextCustom": {
+ "message": "Personalizzato"
}
}
diff --git a/src/public/_locales/ja/messages.json b/src/public/_locales/ja/messages.json
index ced8c60..347ad7b 100644
--- a/src/public/_locales/ja/messages.json
+++ b/src/public/_locales/ja/messages.json
@@ -7,5 +7,20 @@
},
"openSidePanelToChat": {
"message": "チャットするためにCopilotを開く"
- }
+ },
+ "contextSummarize": {
+ "message": "要約"
+ },
+ "contextExplain": {
+ "message": "説明"
+ },
+ "contextRephrase": {
+ "message": "言い換え"
+ },
+ "contextTranslate" :{
+ "message": "翻訳"
+ },
+ "contextCustom": {
+ "message": "カスタム"
+ }
}
\ No newline at end of file
diff --git a/src/public/_locales/ml/messages.json b/src/public/_locales/ml/messages.json
index ad09cde..4a6699a 100644
--- a/src/public/_locales/ml/messages.json
+++ b/src/public/_locales/ml/messages.json
@@ -7,5 +7,20 @@
},
"openSidePanelToChat": {
"message": "ചാറ്റ് ചെയ്യാന് സൈഡ് പാനല് തുറക്കുക"
+ },
+ "contextSummarize": {
+ "message": "സംഗ്രഹിക്കുക"
+ },
+ "contextExplain": {
+ "message": "വിശദീകരിക്കുക"
+ },
+ "contextRephrase": {
+ "message": "പുനഃരൂപീകരിക്കുക"
+ },
+ "contextTranslate" :{
+ "message": "വിവർത്തനം ചെയ്യുക"
+ },
+ "contextCustom": {
+ "message": "ഇഷ്ടാനുസൃതം"
}
}
\ No newline at end of file
diff --git a/src/public/_locales/ru/messages.json b/src/public/_locales/ru/messages.json
index 94b1dfa..61c6a1d 100644
--- a/src/public/_locales/ru/messages.json
+++ b/src/public/_locales/ru/messages.json
@@ -7,5 +7,20 @@
},
"openSidePanelToChat": {
"message": "Открыть Copilot для чата"
+ },
+ "contextSummarize": {
+ "message": "Обобщить"
+ },
+ "contextExplain": {
+ "message": "Объяснить"
+ },
+ "contextRephrase": {
+ "message": "Перефразировать"
+ },
+ "contextTranslate" :{
+ "message": "Перевести"
+ },
+ "contextCustom": {
+ "message": "Пользовательский"
}
}
\ No newline at end of file
diff --git a/src/public/_locales/zh_CN/messages.json b/src/public/_locales/zh_CN/messages.json
index 49dadd1..e9be275 100644
--- a/src/public/_locales/zh_CN/messages.json
+++ b/src/public/_locales/zh_CN/messages.json
@@ -7,5 +7,20 @@
},
"openSidePanelToChat": {
"message": "打开Copilot进行聊天"
+ },
+ "contextSummarize": {
+ "message": "总结"
+ },
+ "contextExplain": {
+ "message": "解释"
+ },
+ "contextRephrase": {
+ "message": "改述"
+ },
+ "contextTranslate" :{
+ "message": "翻译"
+ },
+ "contextCustom": {
+ "message": "自定义"
}
}
\ No newline at end of file
diff --git a/src/services/app.ts b/src/services/app.ts
index 72cc441..bbae388 100644
--- a/src/services/app.ts
+++ b/src/services/app.ts
@@ -108,4 +108,13 @@ export const getTotalFilePerKB = async (): Promise => {
export const setTotalFilePerKB = async (totalFilePerKB: number): Promise => {
await storage.set("totalFilePerKB", totalFilePerKB);
-};
\ No newline at end of file
+};
+
+export const getNoOfRetrievedDocs = async (): Promise => {
+ const noOfRetrievedDocs = await storage.get("noOfRetrievedDocs");
+ return noOfRetrievedDocs || 4
+}
+
+export const setNoOfRetrievedDocs = async (noOfRetrievedDocs: number): Promise => {
+ await storage.set("noOfRetrievedDocs", noOfRetrievedDocs);
+}
\ No newline at end of file
diff --git a/src/services/ollama.ts b/src/services/ollama.ts
index 6ce83a8..7b0a6a8 100644
--- a/src/services/ollama.ts
+++ b/src/services/ollama.ts
@@ -2,7 +2,7 @@ import { Storage } from "@plasmohq/storage"
import { cleanUrl } from "../libs/clean-url"
import { urlRewriteRuntime } from "../libs/runtime"
import { getChromeAIModel } from "./chrome"
-import { setTotalFilePerKB } from "./app"
+import { setNoOfRetrievedDocs, setTotalFilePerKB } from "./app"
const storage = new Storage()
@@ -326,12 +326,14 @@ export const saveForRag = async (
model: string,
chunkSize: number,
overlap: number,
- totalFilePerKB: number
+ totalFilePerKB: number,
+ noOfRetrievedDocs: number
) => {
await setDefaultEmbeddingModelForRag(model)
await setDefaultEmbeddingChunkSize(chunkSize)
await setDefaultEmbeddingChunkOverlap(overlap)
await setTotalFilePerKB(totalFilePerKB)
+ await setNoOfRetrievedDocs(noOfRetrievedDocs)
}
export const getWebSearchPrompt = async () => {
diff --git a/wxt.config.ts b/wxt.config.ts
index 10b2eff..c323981 100644
--- a/wxt.config.ts
+++ b/wxt.config.ts
@@ -37,7 +37,7 @@ export default defineConfig({
topLevelAwait({
promiseExportName: "__tla",
promiseImportName: (i) => `__tla_${i}`
- })
+ }) as any
],
build: {
rollupOptions: {
@@ -50,7 +50,7 @@ export default defineConfig({
outDir: "build",
manifest: {
- version: "1.2.0",
+ version: "1.2.1",
name:
process.env.TARGET === "firefox"
? "Page Assist - A Web UI for Local AI Models"
@@ -90,4 +90,4 @@ export default defineConfig({
? firefoxMV2Permissions
: chromeMV3Permissions
}
-})
+}) as any
\ No newline at end of file