image-mapper: fix bad aliases in the data

There are images where the alias data in known to be incorrect and it is
non-trivial to fix on our side at the moment. For now, let's correct
those aliases in the mapper to improve the results.

Add an integration test that tests against the actual data returned by
GraphQL so we have a baseline for identifying regressions and codifying
known-good output.

Author: ribbybibby

image-mapper/README.md

@@ -78,3 +78,15 @@ docker run -it --rm image-mapper ghcr.io/stakater/reloader:v1.4.1
 # Or, pass a list of images from a text file
 docker run -i --rm image-mapper -- - < images.txt
 ```
+
+## Development
+
+You can run integration tests against the actual catalog endpoint by setting
+`IMAGE_MAPPER_RUN_INTEGRATION_TESTS=1`:
+
+```
+IMAGE_MAPPER_RUN_INTEGRATION_TESTS=1 go test ./...
+```
+
+This identifies regressions in the mapping logic or the catalog data by
+recording known matches.

image-mapper/internal/mapper/mapper.go

@@ -1,11 +1,8 @@
 package mapper
 
 import (
-	"bytes"
 	"context"
-	"encoding/json"
 	"fmt"
-	"net/http"
 	"slices"
 	"strings"
 
@@ -44,46 +41,6 @@ func NewMapper(ctx context.Context, opts ...Option) (*Mapper, error) {
 	return m, nil
 }
 
-// Repo describes a repo in the catalog
-type Repo struct {
-	Name        string   `json:"name"`
-	CatalogTier string   `json:"catalogTier"`
-	Aliases     []string `json:"aliases"`
-}
-
-func listRepos(ctx context.Context) ([]Repo, error) {
-	c := &http.Client{}
-
-	buf := bytes.NewReader([]byte(`{"query":"query OrganizationImageCatalog($organization: ID!) {\n  repos(filter: {uidp: {childrenOf: $organization}}) {\n    name\n    aliases\n  catalogTier\n  }\n}","variables":{"excludeDates":true,"excludeEpochs":true,"organization":"ce2d1984a010471142503340d670612d63ffb9f6"}}`))
-	req, err := http.NewRequestWithContext(ctx, http.MethodPost, "https://data.chainguard.dev/query?id=PrivateImageCatalog", buf)
-	if err != nil {
-		return nil, fmt.Errorf("constructing request: %w", err)
-	}
-	req.Header.Add("Content-Type", "application/json")
-	req.Header.Add("User-Agent", "image-mapper")
-
-	resp, err := c.Do(req)
-	if err != nil {
-		return nil, fmt.Errorf("making request: %w", err)
-	}
-	defer resp.Body.Close()
-
-	if resp.StatusCode != http.StatusOK {
-		return nil, fmt.Errorf("unexpected status code: %d", resp.StatusCode)
-	}
-
-	var data struct {
-		Data struct {
-			Repos []Repo `json:"repos"`
-		} `json:"data"`
-	}
-	if err := json.NewDecoder(resp.Body).Decode(&data); err != nil {
-		return nil, fmt.Errorf("unmarshaling body: %w", err)
-	}
-
-	return data.Data.Repos, nil
-}
-
 // MapAll returns mappings for all the images returned by the iterator
 func (m *Mapper) MapAll(it Iterator) ([]*Mapping, error) {
 	mapped := make(map[string]struct{})
@@ -145,6 +102,7 @@ func (m *Mapper) Map(image string) (*Mapping, error) {
 	for match := range matches {
 		results = append(results, match)
 	}
+	slices.Sort(results)
 
 	return &Mapping{
 		Image:   image,