Phase 4: upload, LLM extraction, import review flow
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
This commit is contained in:
Jānis Kacēns
@@ -9,6 +9,10 @@ import (
|
||||
"qbank/internal/auth"
|
||||
)
|
||||
|
||||
var tmplFuncs = template.FuncMap{
|
||||
"inc": func(i int) int { return i + 1 },
|
||||
}
|
||||
|
||||
// Renderer parses and executes HTML templates from a directory.
|
||||
type Renderer struct {
|
||||
dir string
|
||||
@@ -18,7 +22,7 @@ func NewRenderer(dir string) *Renderer { return &Renderer{dir: dir} }
|
||||
|
||||
// Render executes layout.html + <name>.html, passing data to the "layout" template.
|
||||
func (r *Renderer) Render(w http.ResponseWriter, status int, name string, data any) {
|
||||
t, err := template.ParseFiles(
|
||||
t, err := template.New("").Funcs(tmplFuncs).ParseFiles(
|
||||
filepath.Join(r.dir, "layout.html"),
|
||||
filepath.Join(r.dir, name+".html"),
|
||||
)
|
||||
|
||||
@@ -0,0 +1,233 @@
|
||||
package handlers
|
||||
|
||||
import (
|
||||
"bytes"
|
||||
"errors"
|
||||
"fmt"
|
||||
"io"
|
||||
"log/slog"
|
||||
"net/http"
|
||||
"os"
|
||||
"path/filepath"
|
||||
"strings"
|
||||
"time"
|
||||
|
||||
"github.com/go-chi/chi/v5"
|
||||
|
||||
"qbank/internal/auth"
|
||||
"qbank/internal/db"
|
||||
"qbank/internal/llm"
|
||||
"qbank/internal/models"
|
||||
"qbank/internal/parse"
|
||||
)
|
||||
|
||||
type UploadHandler struct {
|
||||
auth *auth.Manager
|
||||
repo *db.Repo
|
||||
llm *llm.Client
|
||||
render *Renderer
|
||||
dataDir string
|
||||
}
|
||||
|
||||
func NewUploadHandler(a *auth.Manager, repo *db.Repo, llmClient *llm.Client, r *Renderer, dataDir string) *UploadHandler {
|
||||
return &UploadHandler{auth: a, repo: repo, llm: llmClient, render: r, dataDir: dataDir}
|
||||
}
|
||||
|
||||
func (h *UploadHandler) UploadGet(w http.ResponseWriter, r *http.Request) {
|
||||
data := BaseData(h.auth, r)
|
||||
h.render.Render(w, http.StatusOK, "upload", data)
|
||||
}
|
||||
|
||||
func (h *UploadHandler) UploadPost(w http.ResponseWriter, r *http.Request) {
|
||||
r.Body = http.MaxBytesReader(w, r.Body, 20<<20) // 20 MB
|
||||
|
||||
renderErr := func(msg string) {
|
||||
data := BaseData(h.auth, r)
|
||||
data["Error"] = msg
|
||||
h.render.Render(w, http.StatusUnprocessableEntity, "upload", data)
|
||||
}
|
||||
|
||||
if !h.auth.CheckCSRF(r) {
|
||||
HTTPError(w, http.StatusForbidden)
|
||||
return
|
||||
}
|
||||
if err := r.ParseMultipartForm(20 << 20); err != nil {
|
||||
renderErr("File too large (max 20 MB).")
|
||||
return
|
||||
}
|
||||
|
||||
file, header, err := r.FormFile("file")
|
||||
if err != nil {
|
||||
renderErr("No file selected.")
|
||||
return
|
||||
}
|
||||
defer file.Close()
|
||||
|
||||
ext := strings.ToLower(filepath.Ext(header.Filename))
|
||||
if ext != ".pdf" && ext != ".docx" {
|
||||
renderErr("Unsupported file type. Upload a .pdf or .docx file.")
|
||||
return
|
||||
}
|
||||
|
||||
// Read file once; use for both saving and extraction.
|
||||
data, err := io.ReadAll(file)
|
||||
if err != nil {
|
||||
renderErr("Failed to read uploaded file.")
|
||||
return
|
||||
}
|
||||
|
||||
// Save to disk.
|
||||
uploadsDir := filepath.Join(h.dataDir, "uploads")
|
||||
if err := os.MkdirAll(uploadsDir, 0755); err != nil {
|
||||
renderErr("Server error: could not create uploads directory.")
|
||||
return
|
||||
}
|
||||
safeName := filepath.Base(header.Filename)
|
||||
savedPath := filepath.Join(uploadsDir, fmt.Sprintf("%d_%s", time.Now().UnixMilli(), safeName))
|
||||
if err := os.WriteFile(savedPath, data, 0644); err != nil {
|
||||
renderErr("Server error: could not save file.")
|
||||
return
|
||||
}
|
||||
|
||||
// Extract text.
|
||||
var text string
|
||||
switch ext {
|
||||
case ".pdf":
|
||||
text, err = parse.ExtractPDF(bytes.NewReader(data))
|
||||
if errors.Is(err, parse.ErrScanPDF) {
|
||||
renderErr("This PDF appears to be image-only (scan-based). Please convert it to a text PDF first.")
|
||||
return
|
||||
}
|
||||
case ".docx":
|
||||
text, err = parse.ExtractDOCX(bytes.NewReader(data))
|
||||
}
|
||||
if err != nil {
|
||||
renderErr("Could not extract text from the document: " + err.Error())
|
||||
return
|
||||
}
|
||||
|
||||
// Chunk and call LLM.
|
||||
chunks := parse.Chunk(text, 10_000)
|
||||
seen := make(map[string]bool)
|
||||
var draftQs []models.DraftQuestion
|
||||
|
||||
for _, chunk := range chunks {
|
||||
qs, err := h.llm.ExtractQuestions(r.Context(), chunk)
|
||||
if err != nil {
|
||||
slog.Warn("llm chunk failed", "err", err)
|
||||
continue
|
||||
}
|
||||
for _, q := range qs {
|
||||
key := db.QuestionID(q.Question)
|
||||
if seen[key] {
|
||||
continue
|
||||
}
|
||||
seen[key] = true
|
||||
dq := models.DraftQuestion{Text: q.Question}
|
||||
for _, a := range q.Answers {
|
||||
dq.Answers = append(dq.Answers, models.DraftAnswer{Text: a.Text, IsCorrect: a.Correct})
|
||||
}
|
||||
draftQs = append(draftQs, dq)
|
||||
}
|
||||
}
|
||||
|
||||
source := strings.TrimSpace(r.FormValue("source"))
|
||||
if source == "" {
|
||||
source = strings.TrimSuffix(safeName, ext)
|
||||
}
|
||||
|
||||
userID := auth.UserFromCtx(r.Context()).ID
|
||||
draftID, err := h.repo.CreateDraft(userID, source, draftQs)
|
||||
if err != nil {
|
||||
renderErr("Server error: could not create import draft.")
|
||||
return
|
||||
}
|
||||
|
||||
http.Redirect(w, r, "/import/"+draftID, http.StatusSeeOther)
|
||||
}
|
||||
|
||||
func (h *UploadHandler) ImportGet(w http.ResponseWriter, r *http.Request) {
|
||||
draftID := chi.URLParam(r, "id")
|
||||
userID := auth.UserFromCtx(r.Context()).ID
|
||||
|
||||
draft, err := h.repo.GetDraftForUser(draftID, userID)
|
||||
if err != nil {
|
||||
HTTPError(w, http.StatusNotFound)
|
||||
return
|
||||
}
|
||||
|
||||
data := BaseData(h.auth, r)
|
||||
data["Draft"] = draft
|
||||
h.render.Render(w, http.StatusOK, "import", data)
|
||||
}
|
||||
|
||||
func (h *UploadHandler) ImportPost(w http.ResponseWriter, r *http.Request) {
|
||||
draftID := chi.URLParam(r, "id")
|
||||
userID := auth.UserFromCtx(r.Context()).ID
|
||||
|
||||
if !h.auth.CheckCSRF(r) {
|
||||
HTTPError(w, http.StatusForbidden)
|
||||
return
|
||||
}
|
||||
|
||||
draft, err := h.repo.GetDraftForUser(draftID, userID)
|
||||
if err != nil {
|
||||
HTTPError(w, http.StatusNotFound)
|
||||
return
|
||||
}
|
||||
|
||||
source := strings.TrimSpace(r.FormValue("source"))
|
||||
|
||||
var imported, skipped int
|
||||
for i, dq := range draft.Questions {
|
||||
if r.FormValue(fmt.Sprintf("delete_%d", i)) == "on" {
|
||||
skipped++
|
||||
continue
|
||||
}
|
||||
|
||||
text := strings.TrimSpace(r.FormValue(fmt.Sprintf("q_text_%d", i)))
|
||||
if text == "" {
|
||||
skipped++
|
||||
continue
|
||||
}
|
||||
|
||||
correctIdx := r.FormValue(fmt.Sprintf("correct_%d", i))
|
||||
var answers []*models.Answer
|
||||
for j := range dq.Answers {
|
||||
aText := strings.TrimSpace(r.FormValue(fmt.Sprintf("a_text_%d_%d", i, j)))
|
||||
if aText == "" {
|
||||
continue
|
||||
}
|
||||
answers = append(answers, &models.Answer{
|
||||
Text: aText,
|
||||
IsCorrect: fmt.Sprintf("%d", j) == correctIdx,
|
||||
})
|
||||
}
|
||||
|
||||
var nCorrect int
|
||||
for _, a := range answers {
|
||||
if a.IsCorrect {
|
||||
nCorrect++
|
||||
}
|
||||
}
|
||||
if len(answers) < 2 || nCorrect != 1 {
|
||||
skipped++
|
||||
continue
|
||||
}
|
||||
|
||||
q := &models.Question{Text: text, Source: source}
|
||||
if err := h.repo.InsertQuestion(q, answers); err != nil {
|
||||
slog.Error("insert question", "err", err)
|
||||
skipped++
|
||||
continue
|
||||
}
|
||||
imported++
|
||||
}
|
||||
|
||||
if err := h.repo.DeleteDraft(draftID); err != nil {
|
||||
slog.Error("delete draft", "id", draftID, "err", err)
|
||||
}
|
||||
|
||||
h.auth.SetFlash(r, fmt.Sprintf("Imported %d question(s), %d skipped.", imported, skipped))
|
||||
http.Redirect(w, r, "/", http.StatusSeeOther)
|
||||
}
|
||||
Reference in New Issue
Block a user