Agent-System/cmd/document-revision-backfill/main.go

package main

import (
	"context"
	"encoding/json"
	"flag"
	"fmt"
	"os"
	"path/filepath"
	"strings"
	"time"

	"chat-agent-system/internal/docrevision"

	"github.com/jackc/pgx/v5/pgxpool"
)

// document-revision-backfill dry-runs or applies controlled document output.md into the ledger.
func main() {
	dsn := flag.String("dsn", os.Getenv("TEST_DATABASE_URL"), "postgres DSN")
	repoRoot := flag.String("repo-root", ".", "repository root for relative work_dir")
	apply := flag.Bool("apply", false, "apply changes (default dry-run)")
	outPath := flag.String("report", "", "write JSON report path")
	flag.Parse()
	if *dsn == "" {
		fmt.Fprintln(os.Stderr, "dsn or TEST_DATABASE_URL required")
		os.Exit(2)
	}
	ctx, cancel := context.WithTimeout(context.Background(), 5*time.Minute)
	defer cancel()
	pool, err := pgxpool.New(ctx, *dsn)
	if err != nil {
		fmt.Fprintln(os.Stderr, err)
		os.Exit(1)
	}
	defer pool.Close()

	rows, err := pool.Query(ctx, `
		SELECT doc_id, COALESCE(work_dir,''), COALESCE(content_sha256,''), COALESCE(current_state,'drafting'), project_id
		FROM document_records
		ORDER BY doc_id`)
	if err != nil {
		fmt.Fprintln(os.Stderr, err)
		os.Exit(1)
	}
	defer rows.Close()

	type item struct {
		DocID   string `json:"doc_id"`
		Status  string `json:"status"`
		Detail  string `json:"detail,omitempty"`
		SHA     string `json:"sha,omitempty"`
		FileSHA string `json:"file_sha,omitempty"`
	}
	report := struct {
		DryRun   bool   `json:"dry_run"`
		Items    []item `json:"items"`
		Success  int    `json:"success"`
		Missing  int    `json:"missing"`
		Mismatch int    `json:"hash_mismatch"`
	}{DryRun: !*apply}

	svc := docrevision.NewService(pool)
	for rows.Next() {
		var docID, workDir, dbSHA, state string
		var projectID int64
		if err := rows.Scan(&docID, &workDir, &dbSHA, &state, &projectID); err != nil {
			fmt.Fprintln(os.Stderr, err)
			os.Exit(1)
		}
		it := item{DocID: docID, SHA: dbSHA}
		if workDir == "" {
			it.Status = "missing"
			it.Detail = "empty work_dir"
			report.Missing++
			report.Items = append(report.Items, it)
			continue
		}
		path := workDir
		if !filepath.IsAbs(path) {
			path = filepath.Join(*repoRoot, workDir)
		}
		body, err := os.ReadFile(filepath.Join(path, "output.md"))
		if err != nil {
			it.Status = "missing"
			it.Detail = err.Error()
			report.Missing++
			report.Items = append(report.Items, it)
			continue
		}
		canon := docrevision.CanonicalizeMarkdown(string(body))
		fileSHA := docrevision.SHA256Hex(canon)
		it.FileSHA = fileSHA
		if dbSHA != "" && !strings.EqualFold(dbSHA, fileSHA) {
			legacy := docrevision.SHA256Hex(docrevision.LegacyDestructiveCanonicalizeMarkdown(string(body)))
			if strings.EqualFold(dbSHA, legacy) {
				it.Status = "legacy_canonicalization_drift"
				it.Detail = "LEGACY_CANONICALIZATION_DRIFT: refuse silent overwrite; use document-content-reconcile"
			} else {
				it.Status = "hash_mismatch"
				it.Detail = "CONTENT_FORK: document_records.content_sha256 != file"
			}
			report.Mismatch++
			report.Items = append(report.Items, it)
			continue
		}
		if *apply {
			if _, err := svc.OpenTemplateVersion(ctx, docID, canon, docrevision.ActorSnapshot{DisplayName: "backfill"}, projectID, workDir); err != nil {
				it.Status = "error"
				it.Detail = err.Error()
				report.Items = append(report.Items, it)
				continue
			}
		}
		it.Status = "success"
		report.Success++
		report.Items = append(report.Items, it)
	}
	if err := rows.Err(); err != nil {
		fmt.Fprintln(os.Stderr, err)
		os.Exit(1)
	}
	enc, _ := json.MarshalIndent(report, "", "  ")
	if *outPath != "" {
		_ = os.WriteFile(*outPath, enc, 0o644)
	}
	fmt.Println(string(enc))
	if report.Mismatch > 0 {
		os.Exit(3)
	}
}