Mirrors/Stirling-PDF
Mirrors/Stirling-PDF
1
0
Fork 0
mirror of https://github.com/Frooodle/Stirling-PDF.git synced 2025-11-16 01:21:16 +01:00
Code Issues Packages Projects Releases Wiki Activity

improving context

Browse Source
This commit is contained in:
Dario Ghunney Ware 2025-11-14 17:58:12 +00:00
parent c37707d9ad
commit 9ffbede49a
8 changed files with 209 additions and 88 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
app/proprietary/src/main/java/stirling/software/proprietary/config/ChatbotVectorStoreConfig.java
View File

@ -28,10 +28,12 @@ public class ChatbotVectorStoreConfig {
public VectorStore chatbotVectorStore(
ObjectProvider<JedisPooled> jedisProvider, EmbeddingModel embeddingModel) {
JedisPooled jedis = jedisProvider.getIfAvailable();
if (jedis != null) {
try {
jedis.ping();
log.info("Initialising Redis vector store for chatbot usage");
return RedisVectorStore.builder(jedis, embeddingModel)
.indexName(DEFAULT_INDEX)
.prefix(DEFAULT_PREFIX)
@ -45,6 +47,7 @@ public class ChatbotVectorStoreConfig {
} else {
log.info("No Redis connection detected; using SimpleVectorStore for chatbot.");
}
return SimpleVectorStore.builder(embeddingModel).build();
}

2
app/proprietary/src/main/java/stirling/software/proprietary/controller/ChatbotController.java
View File

@ -107,7 +107,7 @@ public class ChatbotController {
warnings.add("Images detected - Images are not currently supported.");
}
warnings.add("Only extracted text is sent for analysis.");
warnings.add("Images are not yet supported. Only extracted text is sent for analysis.");
if (session != null && session.isOcrRequested()) {
warnings.add("OCR requested – uses credits .");
}

49
app/proprietary/src/main/java/stirling/software/proprietary/service/chatbot/ChatbotContextCompressor.java Normal file
View File

@ -0,0 +1,49 @@
package stirling.software.proprietary.service.chatbot;
import java.util.List;
import org.springframework.ai.document.Document;
import org.springframework.stereotype.Component;
import org.springframework.util.CollectionUtils;
@Component
public class ChatbotContextCompressor {
private static final int DEFAULT_SUMMARY_LIMIT = 3000;
private static final int MIN_CHUNK_SNIPPET = 160;
public String summarize(List<Document> documents, int requestedLimit) {
if (CollectionUtils.isEmpty(documents)) {
return "No contextual snippets available for this session.";
}
int maxChars =
requestedLimit > 0
? Math.min(requestedLimit, DEFAULT_SUMMARY_LIMIT)
: DEFAULT_SUMMARY_LIMIT;
StringBuilder builder = new StringBuilder();
int perChunkLimit = Math.max(MIN_CHUNK_SNIPPET, maxChars / Math.max(documents.size(), 1));
for (Document doc : documents) {
if (builder.length() >= maxChars) {
break;
}
String chunkOrder = doc.getMetadata().getOrDefault("chunkOrder", "?").toString();
String text = trimContent(doc.getText(), perChunkLimit);
builder.append("Chunk ").append(chunkOrder).append(": ").append(text).append('\n');
}
if (builder.length() == 0) {
return "Unable to summarise context; original content unavailable.";
}
return builder.substring(0, Math.min(builder.length(), maxChars)).trim();
}
private String trimContent(String content, int perChunkLimit) {
if (content == null || content.isBlank()) {
return "(empty chunk)";
}
String normalized = content.replaceAll("\\s+", " ").trim();
if (normalized.length() <= perChunkLimit) {
return normalized;
}
return normalized.substring(0, Math.max(0, perChunkLimit - 3)) + "...";
}
}

51
app/proprietary/src/main/java/stirling/software/proprietary/service/chatbot/ChatbotConversationService.java
View File

@ -47,6 +47,8 @@ public class ChatbotConversationService {
private final ChatbotCacheService cacheService;
private final ChatbotFeatureProperties featureProperties;
private final ChatbotRetrievalService retrievalService;
private final ChatbotContextCompressor contextCompressor;
private final ChatbotMemoryService memoryService;
private final ChatbotUsageService usageService;
private final ObjectMapper objectMapper;
private final AtomicBoolean modelSwitchVerified = new AtomicBoolean(false);
@ -79,6 +81,9 @@ public class ChatbotConversationService {
List<Document> context =
retrievalService.retrieveTopK(
request.getSessionId(), request.getPrompt(), settings);
String contextSummary =
contextCompressor.summarize(
context, (int) Math.max(settings.maxPromptCharacters() / 2, 1000));
ModelReply nanoReply =
invokeModel(
@ -87,6 +92,7 @@ public class ChatbotConversationService {
request.getPrompt(),
session,
context,
contextSummary,
cacheEntry.getMetadata());
boolean shouldEscalate =
@ -106,6 +112,7 @@ public class ChatbotConversationService {
request.getPrompt(),
session,
context,
contextSummary,
cacheEntry.getMetadata());
}
@ -116,6 +123,8 @@ public class ChatbotConversationService {
finalReply.completionTokens());
session.setUsageSummary(usageSummary);
memoryService.recordTurn(session, request.getPrompt(), finalReply.answer());
return ChatbotResponse.builder()
.sessionId(request.getSessionId())
.modelUsed(
@ -200,8 +209,10 @@ public class ChatbotConversationService {
String prompt,
ChatbotSession session,
List<Document> context,
String contextSummary,
Map<String, String> metadata) {
Prompt requestPrompt = buildPrompt(settings, model, prompt, session, context, metadata);
Prompt requestPrompt =
buildPrompt(settings, model, prompt, session, context, contextSummary, metadata);
ChatResponse response;
try {
response = chatModel.call(requestPrompt);
@ -244,16 +255,9 @@ public class ChatbotConversationService {
String question,
ChatbotSession session,
List<Document> context,
String contextSummary,
Map<String, String> metadata) {
StringBuilder contextBuilder = new StringBuilder();
for (Document chunk : context) {
contextBuilder
.append("[Chunk ")
.append(chunk.getMetadata().getOrDefault("chunkOrder", "?"))
.append("]\n")
.append(chunk.getText())
.append("\n\n");
}
String chunkOutline = buildChunkOutline(context);
String metadataSummary =
metadata.entrySet().stream()
.map(entry -> entry.getKey() + ": " + entry.getValue())
@ -277,8 +281,10 @@ public class ChatbotConversationService {
+ session.isOcrRequested()
+ "\n"
+ imageDirective
+ "\nContext:\n"
+ contextBuilder
+ "\nContext summary:\n"
+ contextSummary
+ "\nContext outline:\n"
+ chunkOutline
+ "Question: "
+ question;
@ -298,6 +304,27 @@ public class ChatbotConversationService {
return builder.build();
}
private String buildChunkOutline(List<Document> context) {
if (context == null || context.isEmpty()) {
return "No chunks retrieved for this question.";
}
StringBuilder outline = new StringBuilder();
for (Document chunk : context) {
String order = chunk.getMetadata().getOrDefault("chunkOrder", "?").toString();
String snippet = chunk.getText();
if (snippet != null) {
snippet = snippet.replaceAll("\\s+", " ").trim();
if (snippet.length() > 240) {
snippet = snippet.substring(0, 237) + "...";
}
} else {
snippet = "(empty)";
}
outline.append("- Chunk ").append(order).append(": ").append(snippet).append("\n");
}
return outline.toString();
}
private ModelReply parseModelResponse(
String raw, long promptTokens, long completionTokens, long totalTokens) {
if (!StringUtils.hasText(raw)) {

52
app/proprietary/src/main/java/stirling/software/proprietary/service/chatbot/ChatbotMemoryService.java Normal file
View File

@ -0,0 +1,52 @@
package stirling.software.proprietary.service.chatbot;
import java.time.Instant;
import java.util.HashMap;
import java.util.List;
import java.util.Map;
import org.springframework.ai.document.Document;
import org.springframework.ai.vectorstore.VectorStore;
import org.springframework.stereotype.Service;
import org.springframework.util.StringUtils;
import lombok.RequiredArgsConstructor;
import lombok.extern.slf4j.Slf4j;
import stirling.software.proprietary.model.chatbot.ChatbotSession;
@Slf4j
@Service
@RequiredArgsConstructor
public class ChatbotMemoryService {
private final VectorStore vectorStore;
public void recordTurn(ChatbotSession session, String prompt, String answer) {
if (session == null) {
return;
}
if (!StringUtils.hasText(prompt) && !StringUtils.hasText(answer)) {
return;
}
Map<String, Object> metadata = new HashMap<>();
metadata.put("sessionId", session.getSessionId());
metadata.put("documentId", session.getDocumentId());
metadata.put("turnType", "conversation");
metadata.put("turnTimestamp", Instant.now().toString());
metadata.put("userId", session.getUserId());
StringBuilder contentBuilder = new StringBuilder();
if (StringUtils.hasText(prompt)) {
contentBuilder.append("User: ").append(prompt.trim()).append("\n");
}
if (StringUtils.hasText(answer)) {
contentBuilder.append("Assistant: ").append(answer.trim());
}
try {
vectorStore.add(List.of(new Document(contentBuilder.toString(), metadata)));
} catch (RuntimeException ex) {
log.warn("Failed to persist chatbot conversation turn: {}", ex.getMessage());
}
}
}

13
frontend/public/locales/en-GB/translation.json
View File

@ -5145,8 +5145,7 @@
"title": "Stirling PDF Bot",
"alphaBadge": "Alpha",
"alphaTitle": "Experimental feature",
"alphaDescription": "This Chatbot feature is in currently in alpha and is subject to change. Image-based content is not supported yet. Responses may be imperfect, so double-check important answers.",
"acceptAlphaLabel": "I understand this feature is experimental and image content is not supported yet.",
"alphaDescription": "Chatbot is in currently in alpha and is subject to change. Responses may be imperfect, please check responses.",
"fileLabel": "Document to query",
"filePlaceholder": "Select an uploaded PDF",
"noFiles": "Upload a PDF from File Manager to start chatting.",
@ -5193,6 +5192,14 @@
"noTextTitle": "No text detected in this PDF",
"noTextBody": "We could not find selectable text in this document. Would you like to run OCR to convert scanned pages into text?",
"noTextDismiss": "Maybe later",
"noTextRunOcr": "Run OCR and retry"
"noTextRunOcr": "Run OCR and retry",
"usage": {
"limitReachedTitle": "Chatbot limit reached",
"limitReachedBody": "You have exceeded the current monthly allocation for the chatbot. Further responses may be throttled.",
"nearingLimitTitle": "Approaching usage limit",
"nearingLimitBody": "You are nearing your monthly chatbot allocation. Consider limiting very large requests."
},
"autoSyncInfo": "Selected documents are synced automatically when the chatbot opens.",
"autoSyncPrompt": "Acknowledge the alpha notice to start syncing automatically."
}
}

123
frontend/src/core/components/chatbot/ChatbotDrawer.tsx
View File

@ -1,6 +1,5 @@
import { useEffect, useLayoutEffect, useMemo, useRef, useState, type KeyboardEvent } from 'react';
import {
ActionIcon,
Badge,
Box,
Button,
@ -13,14 +12,12 @@ import {
Switch,
Text,
Textarea,
Tooltip,
} from '@mantine/core';
import { useMediaQuery, useViewportSize } from '@mantine/hooks';
import { useTranslation } from 'react-i18next';
import SmartToyRoundedIcon from '@mui/icons-material/SmartToyRounded';
import WarningAmberRoundedIcon from '@mui/icons-material/WarningAmberRounded';
import SendRoundedIcon from '@mui/icons-material/SendRounded';
import RefreshRoundedIcon from '@mui/icons-material/RefreshRounded';
import CloseRoundedIcon from '@mui/icons-material/CloseRounded';
import { useChatbot } from '@app/contexts/ChatbotContext';
@ -47,8 +44,6 @@ interface ChatMessage {
createdAt: Date;
}
const ALPHA_ACK_KEY = 'stirling.chatbot.alphaAck';
function createMessageId() {
if (typeof crypto !== 'undefined' && crypto.randomUUID) {
return crypto.randomUUID();
@ -68,7 +63,6 @@ const ChatbotDrawer = () => {
const { show } = useToast();
const files = selectors.getFiles();
const [selectedFileId, setSelectedFileId] = useState<string | undefined>();
const [alphaAccepted, setAlphaAccepted] = useState(false);
const [runOcr, setRunOcr] = useState(false);
const [isStartingSession, setIsStartingSession] = useState(false);
const [isSendingMessage, setIsSendingMessage] = useState(false);
@ -94,11 +88,6 @@ const ChatbotDrawer = () => {
return;
}
const storedAck = typeof window !== 'undefined'
? window.localStorage.getItem(ALPHA_ACK_KEY) === 'true'
: false;
setAlphaAccepted(storedAck);
if (preferredFileId) {
setSelectedFileId(preferredFileId);
setPreferredFileId(undefined);
@ -161,6 +150,8 @@ const ChatbotDrawer = () => {
setContextStats(null);
setMessages([]);
setWarnings([]);
setPendingOcrRetry(false);
setNoTextModalOpen(false);
}
}, [sessionInfo, selectedFileId]);
@ -192,17 +183,6 @@ const ChatbotDrawer = () => {
};
}, [isMobile, isOpen, sidebarRefs.toolPanelRef]);
const handleAlphaAccept = (checked: boolean) => {
setAlphaAccepted(checked);
if (typeof window !== 'undefined') {
if (checked) {
window.localStorage.setItem(ALPHA_ACK_KEY, 'true');
} else {
window.localStorage.removeItem(ALPHA_ACK_KEY);
}
}
};
const withStatus = async <T,>(label: string, fn: () => Promise<T>): Promise<T> => {
setStatusMessage(label);
try {
@ -228,14 +208,6 @@ const ChatbotDrawer = () => {
if (!ensureFileSelected() || !selectedFile) {
return;
}
if (!alphaAccepted) {
show({
alertType: 'neutral',
title: t('chatbot.toasts.ackTitle', 'Accept alpha notice'),
body: t('chatbot.toasts.ackBody', 'Please acknowledge the alpha warning before starting.'),
});
return;
}
setIsStartingSession(true);
try {
let workingFile: File = selectedFile;
@ -271,7 +243,7 @@ const ChatbotDrawer = () => {
text: extractionResult.text,
metadata,
ocrRequested: shouldRunOcr,
warningsAccepted: alphaAccepted,
warningsAccepted: true,
};
const response = await withStatus(
@ -302,6 +274,36 @@ const ChatbotDrawer = () => {
}
};
useEffect(() => {
if (
!isOpen ||
!selectedFile ||
sessionInfo ||
isStartingSession ||
pendingOcrRetry ||
noTextModalOpen
) {
return;
}
let cancelled = false;
handleSessionStart().catch((error) => {
if (!cancelled) {
console.error('[Chatbot] Auto-sync failed', error);
}
});
return () => {
cancelled = true;
};
}, [isOpen, selectedFile, sessionInfo, isStartingSession, pendingOcrRetry, noTextModalOpen, runOcr]);
useEffect(() => {
if (!sessionInfo) {
return;
}
setSessionInfo(null);
setContextStats(null);
}, [runOcr]);
const handleSendMessage = async () => {
if (!sessionInfo) {
show({
@ -495,25 +497,6 @@ const ChatbotDrawer = () => {
transitionProps={{ transition: 'slide-left', duration: 200 }}
>
<Stack gap="sm" h="100%" style={{ minHeight: 0 }}>
<Box
p="sm"
style={{
border: '1px solid var(--border-subtle)',
borderRadius: 8,
backgroundColor: 'var(--bg-subtle)',
display: 'flex',
gap: '0.5rem',
alignItems: 'flex-start',
}}
>
<WarningAmberRoundedIcon fontSize="small" style={{ color: 'var(--text-warning)' }} />
<Box>
<Text fw={600}>{t('chatbot.alphaTitle', 'Experimental feature')}</Text>
<Text size="sm">
{t('chatbot.alphaDescription', 'This chatbot is in alpha. It currently ignores images and may produce inaccurate answers. Your PDF text stays local until you confirm you want to chat.')}
</Text>
</Box>
</Box>
<Group align="flex-end" justify="space-between" gap="md" wrap="wrap">
<Select
@ -526,11 +509,6 @@ const ChatbotDrawer = () => {
style={{ flex: '1 1 200px' }}
/>
<Stack gap={4} style={{ minWidth: 160 }}>
<Switch
checked={alphaAccepted}
label={t('chatbot.acceptAlphaLabel', 'Alpha notice acknowledged')}
onChange={(event) => handleAlphaAccept(event.currentTarget.checked)}
/>
<Switch
checked={runOcr}
onChange={(event) => setRunOcr(event.currentTarget.checked)}
@ -539,19 +517,6 @@ const ChatbotDrawer = () => {
</Stack>
</Group>
<Button
fullWidth
variant="filled"
leftSection={<RefreshRoundedIcon fontSize="small" />}
loading={isStartingSession}
onClick={() => handleSessionStart()}
disabled={!selectedFile || !alphaAccepted}
>
{sessionInfo
? t('chatbot.refreshButton', 'Re-sync document')
: t('chatbot.startButton', 'Send document to chat')}
</Button>
{statusMessage && (
<Box
p="sm"
@ -596,6 +561,28 @@ const ChatbotDrawer = () => {
</Group>
</Box>
))}
{isOpen && (
<Box
p="sm"
bg="var(--bg-muted)"
style={{ borderRadius: 12, border: '1px solid var(--border-subtle)' }}
>
<Group gap="xs" align="flex-start">
<WarningAmberRoundedIcon fontSize="small" style={{ color: 'var(--text-warning)' }} />
<Box>
<Text size="sm" fw={600}>
{t('chatbot.alphaTitle', 'Experimental feature')}
</Text>
<Text size="sm">
{t(
'chatbot.alphaDescription',
'This chatbot is in alpha. It currently ignores images and may produce inaccurate answers.'
)}
</Text>
</Box>
</Group>
</Box>
)}
{messages.length === 0 && (
<Text size="sm" c="dimmed">
{t('chatbot.emptyState', 'Ask a question about your PDF to start the conversation.')}

4
frontend/src/core/components/viewer/useViewerRightRailButtons.tsx
View File

@ -7,14 +7,10 @@ import LocalIcon from '@app/components/shared/LocalIcon';
import { Tooltip } from '@app/components/shared/Tooltip';
import { SearchInterface } from '@app/components/viewer/SearchInterface';
import ViewerAnnotationControls from '@app/components/shared/rightRail/ViewerAnnotationControls';
import { useFileState } from '@app/contexts/FileContext';
export function useViewerRightRailButtons() {
const { t } = useTranslation();
const viewer = useViewer();
const { selectors } = useFileState();
const filesSignature = selectors.getFilesSignature();
const files = useMemo(() => selectors.getFiles(), [selectors, filesSignature]);
const [isPanning, setIsPanning] = useState<boolean>(() => viewer.getPanState()?.isPanning ?? false);
// Lift i18n labels out of memo for clarity