narratio/internal/artifacts/prepared_input.go

package artifacts

import (
	"crypto/sha256"
	"encoding/hex"
	"errors"
	"fmt"
	"io"
	"path/filepath"
	"strings"

	"gitea.maximumdirect.net/eric/narratio/internal/artifactpolicy"
	"gitea.maximumdirect.net/eric/narratio/internal/fileops"
	"gitea.maximumdirect.net/eric/narratio/internal/manifest"
)

// ErrPreparedInputAbsent reports that the current manifest has no record for a
// supported prepared stable input.
var ErrPreparedInputAbsent = errors.New("prepared input absent")

// PreparedInputAbsentError identifies the prepared source absent from the
// current manifest.
type PreparedInputAbsentError struct {
	SourceID string
}

func (e *PreparedInputAbsentError) Error() string {
	return fmt.Sprintf("%s: %q", ErrPreparedInputAbsent, e.SourceID)
}

func (e *PreparedInputAbsentError) Unwrap() error {
	return ErrPreparedInputAbsent
}

// PreparedInputIdentity is the verified identity of one canonical prepared
// session input.
type PreparedInputIdentity struct {
	SourceID     string
	ManifestKind string
	Path         string
	RelativePath string
	Checksum     string
	Size         int64
}

// ResolvePreparedInput resolves a prepared stable source exclusively from its
// current manifest record and verifies the canonical file's identity.
func ResolvePreparedInput(paths SessionPaths, m *manifest.Manifest, sourceID string) (PreparedInputIdentity, error) {
	descriptor, ok := artifactpolicy.DescribePreparedInputSource(sourceID)
	if !ok {
		return PreparedInputIdentity{}, fmt.Errorf("unsupported prepared input source %q", sourceID)
	}

	rootPath, canonicalPath, relativePath, err := preparedInputCanonicalPaths(paths, descriptor)
	if err != nil {
		return PreparedInputIdentity{}, err
	}

	matching := make([]manifest.InputRecord, 0, 1)
	if m != nil {
		for _, record := range m.Inputs {
			if strings.TrimSpace(record.Kind) == descriptor.ManifestKind {
				matching = append(matching, record)
			}
		}
	}
	if len(matching) == 0 {
		if m != nil {
			for _, record := range m.Inputs {
				recordedPath, pathErr := resolvePreparedManifestPath(paths, record.Path, rootPath)
				if pathErr == nil && recordedPath == canonicalPath {
					return PreparedInputIdentity{}, fmt.Errorf(
						"prepared input source %q canonical path is recorded with manifest kind %q, want %q",
						descriptor.SourceID,
						strings.TrimSpace(record.Kind),
						descriptor.ManifestKind,
					)
				}
			}
		}
		return PreparedInputIdentity{}, &PreparedInputAbsentError{SourceID: descriptor.SourceID}
	}
	if len(matching) != 1 {
		return PreparedInputIdentity{}, fmt.Errorf(
			"prepared input source %q has %d manifest records for kind %q; want exactly one",
			descriptor.SourceID,
			len(matching),
			descriptor.ManifestKind,
		)
	}

	record := matching[0]
	recordedPath, err := resolvePreparedManifestPath(paths, record.Path, rootPath)
	if err != nil {
		return PreparedInputIdentity{}, fmt.Errorf("prepared input source %q manifest path: %w", descriptor.SourceID, err)
	}
	if recordedPath != canonicalPath {
		return PreparedInputIdentity{}, fmt.Errorf(
			"prepared input source %q manifest path %q does not match canonical path %q",
			descriptor.SourceID,
			recordedPath,
			canonicalPath,
		)
	}
	declaredChecksum := strings.TrimSpace(record.Checksum)
	if declaredChecksum == "" {
		return PreparedInputIdentity{}, fmt.Errorf("prepared input source %q manifest checksum is required", descriptor.SourceID)
	}

	file, err := fileops.OpenConfinedRegularFile(rootPath, relativePath)
	if err != nil {
		return PreparedInputIdentity{}, fmt.Errorf("open prepared input source %q: %w", descriptor.SourceID, err)
	}
	digest := sha256.New()
	size, readErr := io.Copy(digest, file)
	closeErr := file.Close()
	if readErr != nil {
		return PreparedInputIdentity{}, fmt.Errorf("checksum prepared input source %q: %w", descriptor.SourceID, readErr)
	}
	if closeErr != nil {
		return PreparedInputIdentity{}, fmt.Errorf("close prepared input source %q: %w", descriptor.SourceID, closeErr)
	}
	if size == 0 {
		return PreparedInputIdentity{}, fmt.Errorf("prepared input source %q is empty", descriptor.SourceID)
	}
	checksum := hex.EncodeToString(digest.Sum(nil))
	if !strings.EqualFold(checksum, declaredChecksum) {
		return PreparedInputIdentity{}, fmt.Errorf(
			"prepared input source %q checksum mismatch: manifest=%q actual=%q",
			descriptor.SourceID,
			declaredChecksum,
			checksum,
		)
	}

	return PreparedInputIdentity{
		SourceID:     descriptor.SourceID,
		ManifestKind: descriptor.ManifestKind,
		Path:         canonicalPath,
		RelativePath: filepath.ToSlash(relativePath),
		Checksum:     checksum,
		Size:         size,
	}, nil
}

func preparedInputCanonicalPaths(
	paths SessionPaths,
	descriptor artifactpolicy.PreparedInputSourceDescriptor,
) (rootPath, canonicalPath, relativePath string, err error) {
	rootPath, err = filepath.Abs(strings.TrimSpace(paths.Root))
	if err != nil || strings.TrimSpace(paths.Root) == "" {
		if err == nil {
			err = fmt.Errorf("session root is required")
		}
		return "", "", "", err
	}
	canonicalPath, err = filepath.Abs(filepath.Join(paths.InputsDir, descriptor.Filename))
	if err != nil {
		return "", "", "", fmt.Errorf("resolve prepared input canonical path: %w", err)
	}
	relativePath, err = filepath.Rel(rootPath, canonicalPath)
	if err != nil || relativePath == "." || relativePath == ".." || strings.HasPrefix(relativePath, ".."+string(filepath.Separator)) {
		if err != nil {
			return "", "", "", fmt.Errorf("resolve prepared input below session root: %w", err)
		}
		return "", "", "", fmt.Errorf("prepared input canonical path %q is outside session root %q", canonicalPath, rootPath)
	}
	return filepath.Clean(rootPath), filepath.Clean(canonicalPath), filepath.Clean(relativePath), nil
}

func resolvePreparedManifestPath(paths SessionPaths, recordedPath, rootPath string) (string, error) {
	if strings.TrimSpace(recordedPath) == "" {
		return "", fmt.Errorf("recorded path is required")
	}
	resolved := ResolveSessionLocalPathForRead(paths, recordedPath)
	if strings.TrimSpace(resolved) == "" {
		return "", fmt.Errorf("recorded path is required")
	}
	absolute, err := filepath.Abs(resolved)
	if err != nil {
		return "", fmt.Errorf("resolve recorded path: %w", err)
	}
	relative, err := filepath.Rel(rootPath, absolute)
	if err != nil {
		return "", fmt.Errorf("resolve recorded path below session root: %w", err)
	}
	if relative == "." || relative == ".." || strings.HasPrefix(relative, ".."+string(filepath.Separator)) {
		return "", fmt.Errorf("recorded path %q is outside session root %q", absolute, rootPath)
	}
	return filepath.Clean(absolute), nil
}