ThothII/tools/tht/internal/pi/latest.go

package pi

import (
	"context"
	"encoding/json"
	"errors"
	"fmt"
	"io"
	"net/http"
	"net/url"
	"os"
	"path/filepath"
	"regexp"
	"strings"
	"time"
)

const (
	piRegistryURL     = "https://registry.npmjs.org"
	registryBodyLimit = 5 << 20
	registryTimeout   = 10 * time.Second
)

var packageNamePattern = regexp.MustCompile(`^(?:@[a-z0-9][a-z0-9._-]*/)?[a-z0-9][a-z0-9._-]*$`)

// RegistryClient resolves a published, stable Pi package version without involving a host npm
// executable. It is deliberately narrow so the update transaction can resolve a version before
// it acquires a lifecycle lock or contacts Docker.
type RegistryClient interface {
	LatestStable(ctx context.Context, packageName string) (string, error)
}

type npmRegistryClient struct {
	client *http.Client
}

// NewNPMRegistryClient returns the bounded HTTPS client for the authoritative npm registry.
func NewNPMRegistryClient() RegistryClient {
	return newNPMRegistryClient(nil)
}

func newNPMRegistryClient(client *http.Client) *npmRegistryClient {
	if client == nil {
		client = &http.Client{Timeout: registryTimeout}
	}
	bounded := *client
	bounded.CheckRedirect = rejectRegistryRedirect
	return &npmRegistryClient{client: &bounded}
}

func rejectRegistryRedirect(request *http.Request, _ []*http.Request) error {
	return fmt.Errorf("Pi registry redirect to %s is not allowed", request.URL.Redacted())
}

// LatestStable returns the greatest released semantic version. npm's dist-tag is not trusted as
// it can point at a prerelease; selecting from the published versions keeps the default stable.
func (c *npmRegistryClient) LatestStable(ctx context.Context, packageName string) (string, error) {
	if !packageNamePattern.MatchString(packageName) {
		return "", errors.New("Pi package name is invalid")
	}
	endpoint, err := registryPackageURL(packageName)
	if err != nil {
		return "", err
	}
	request, err := http.NewRequestWithContext(ctx, http.MethodGet, endpoint, nil)
	if err != nil {
		return "", fmt.Errorf("create Pi registry request: %w", err)
	}
	request.Header.Set("Accept", "application/json")
	response, err := c.client.Do(request)
	if err != nil {
		return "", fmt.Errorf("query Pi registry: %w", err)
	}
	defer response.Body.Close()
	if response.StatusCode == http.StatusNotFound {
		return "", errors.New("Pi package was not found in the registry")
	}
	if response.StatusCode < http.StatusOK || response.StatusCode >= http.StatusMultipleChoices {
		return "", fmt.Errorf("Pi registry returned HTTP %d", response.StatusCode)
	}

	var payload struct {
		Versions map[string]json.RawMessage `json:"versions"`
	}
	body, err := io.ReadAll(io.LimitReader(response.Body, registryBodyLimit+1))
	if err != nil {
		return "", fmt.Errorf("read Pi registry response: %w", err)
	}
	if len(body) > registryBodyLimit {
		return "", errors.New("Pi registry response is too large")
	}
	decoder := json.NewDecoder(strings.NewReader(string(body)))
	if err := decoder.Decode(&payload); err != nil {
		return "", errors.New("Pi registry response is invalid")
	}
	var trailing any
	if err := decoder.Decode(&trailing); err != io.EOF {
		return "", errors.New("Pi registry response is invalid")
	}

	var selected semanticVersion
	found := false
	for rawVersion, rawMetadata := range payload.Versions {
		if !json.Valid(rawMetadata) {
			return "", errors.New("Pi registry response is invalid")
		}
		version, err := parseSemanticVersion(rawVersion)
		if err != nil {
			return "", errors.New("Pi registry response contains an invalid semantic version key")
		}
		if version.prerelease != "" {
			continue
		}
		if !found || selected.less(version) {
			selected, found = version, true
		}
	}
	if !found {
		return "", errors.New("Pi registry contains no stable semantic version")
	}
	return selected.raw, nil
}

func registryPackageURL(packageName string) (string, error) {
	base, err := url.Parse(piRegistryURL)
	if err != nil || base.Scheme != "https" || base.Host != "registry.npmjs.org" {
		return "", errors.New("Pi registry endpoint is not an authoritative HTTPS URL")
	}
	base.Path = "/" + packageName
	base.RawPath = "/" + url.PathEscape(packageName)
	return base.String(), nil
}

// ReadRuntimePackageName obtains the exact Pi package from the dependency manifest copied by
// docker/core.Dockerfile. Version discovery therefore follows the runtime the image actually
// builds, rather than a second, independently maintained package name.
func ReadRuntimePackageName(projectDirectory string) (string, error) {
	contents, err := os.ReadFile(filepath.Join(projectDirectory, "docker", "pi-runtime", "package.json"))
	if err != nil {
		return "", fmt.Errorf("read Pi runtime package manifest: %w", err)
	}
	var manifest struct {
		Dependencies map[string]string `json:"dependencies"`
	}
	if err := json.Unmarshal(contents, &manifest); err != nil {
		return "", errors.New("Pi runtime package manifest is invalid")
	}
	const piDependency = "@earendil-works/pi-coding-agent"
	if !packageNamePattern.MatchString(piDependency) || strings.TrimSpace(manifest.Dependencies[piDependency]) == "" {
		return "", errors.New("Pi runtime package manifest does not declare the Pi dependency")
	}
	return piDependency, nil
}

// ResolveRequestedVersion validates an explicit version without a lookup, or resolves the latest
// published stable version before the caller starts any lifecycle mutation.
func ResolveRequestedVersion(ctx context.Context, requested string, packageName string, registry RegistryClient) (string, error) {
	if strings.TrimSpace(requested) != "" {
		version, err := parseSemanticVersion(requested)
		if err != nil {
			return "", err
		}
		return version.raw, nil
	}
	if registry == nil {
		return "", errors.New("Pi registry client is unavailable")
	}
	version, err := registry.LatestStable(ctx, packageName)
	if err != nil {
		return "", err
	}
	parsed, err := parseSemanticVersion(version)
	if err != nil || parsed.prerelease != "" {
		return "", errors.New("Pi registry did not return a stable semantic version")
	}
	return parsed.raw, nil
}

type semanticVersion struct {
	raw        string
	major      string
	minor      string
	patch      string
	prerelease string
}

var semanticVersionPattern = regexp.MustCompile(`^(0|[1-9][0-9]*)\.(0|[1-9][0-9]*)\.(0|[1-9][0-9]*)(?:-([0-9A-Za-z-]+(?:\.[0-9A-Za-z-]+)*))?(?:\+[0-9A-Za-z-]+(?:\.[0-9A-Za-z-]+)*)?$`)

func parseSemanticVersion(raw string) (semanticVersion, error) {
	match := semanticVersionPattern.FindStringSubmatch(raw)
	if len(match) == 0 {
		return semanticVersion{}, errors.New("Pi version must be a strict semantic version")
	}
	if match[4] != "" {
		for _, identifier := range strings.Split(match[4], ".") {
			if allDigits(identifier) && len(identifier) > 1 && identifier[0] == '0' {
				return semanticVersion{}, errors.New("Pi version must be a strict semantic version")
			}
		}
	}
	return semanticVersion{raw: raw, major: match[1], minor: match[2], patch: match[3], prerelease: match[4]}, nil
}

func allDigits(value string) bool {
	for _, character := range value {
		if character < '0' || character > '9' {
			return false
		}
	}
	return value != ""
}

func (left semanticVersion) less(right semanticVersion) bool {
	for _, pair := range [][2]string{{left.major, right.major}, {left.minor, right.minor}, {left.patch, right.patch}} {
		if len(pair[0]) != len(pair[1]) {
			return len(pair[0]) < len(pair[1])
		}
		if pair[0] != pair[1] {
			return pair[0] < pair[1]
		}
	}
	return false
}