rework media processing a little bit (#191)

* rework media processing a little bit

* review changes

internal/ap/extract.go

@@ -384,10 +384,7 @@ func ExtractAttachment(i Attachmentable) (*gtsmodel.MediaAttachment, error) {
 	attachment.RemoteURL = attachmentURL.String()
 
 	mediaType := i.GetActivityStreamsMediaType()
-	if mediaType == nil {
-		return nil, errors.New("no media type")
-	}
-	if mediaType.Get() == "" {
+	if mediaType == nil || mediaType.Get() == "" {
 		return nil, errors.New("no media type")
 	}
 	attachment.File.ContentType = mediaType.Get()

internal/ap/extractattachments_test.go

@@ -0,0 +1,122 @@
+/*
+   GoToSocial
+   Copyright (C) 2021 GoToSocial Authors admin@gotosocial.org
+
+   This program is free software: you can redistribute it and/or modify
+   it under the terms of the GNU Affero General Public License as published by
+   the Free Software Foundation, either version 3 of the License, or
+   (at your option) any later version.
+
+   This program is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+   GNU Affero General Public License for more details.
+
+   You should have received a copy of the GNU Affero General Public License
+   along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+package ap_test
+
+import (
+	"testing"
+
+	"github.com/go-fed/activity/streams"
+	"github.com/go-fed/activity/streams/vocab"
+	"github.com/stretchr/testify/suite"
+	"github.com/superseriousbusiness/gotosocial/internal/ap"
+	"github.com/superseriousbusiness/gotosocial/testrig"
+)
+
+func document1() vocab.ActivityStreamsDocument {
+	document1 := streams.NewActivityStreamsDocument()
+
+	document1MediaType := streams.NewActivityStreamsMediaTypeProperty()
+	document1MediaType.Set("image/jpeg")
+	document1.SetActivityStreamsMediaType(document1MediaType)
+
+	document1URL := streams.NewActivityStreamsUrlProperty()
+	document1URL.AppendIRI(testrig.URLMustParse("https://s3-us-west-2.amazonaws.com/plushcity/media_attachments/files/106/867/380/219/163/828/original/88e8758c5f011439.jpg"))
+	document1.SetActivityStreamsUrl(document1URL)
+
+	document1Name := streams.NewActivityStreamsNameProperty()
+	document1Name.AppendXMLSchemaString("It's a cute plushie.")
+	document1.SetActivityStreamsName(document1Name)
+
+	document1Blurhash := streams.NewTootBlurhashProperty()
+	document1Blurhash.Set("UxQ0EkRP_4tRxtRjWBt7%hozM_ayV@oLf6WB")
+	document1.SetTootBlurhash(document1Blurhash)
+
+	return document1
+}
+
+func attachment1() vocab.ActivityStreamsAttachmentProperty {
+	attachment1 := streams.NewActivityStreamsAttachmentProperty()
+	attachment1.AppendActivityStreamsDocument(document1())
+	return attachment1
+}
+
+type ExtractTestSuite struct {
+	suite.Suite
+}
+
+func (suite *ExtractTestSuite) TestExtractAttachments() {
+	note := streams.NewActivityStreamsNote()
+	note.SetActivityStreamsAttachment(attachment1())
+
+	attachments, err := ap.ExtractAttachments(note)
+	suite.NoError(err)
+	suite.Len(attachments, 1)
+
+	attachment1 := attachments[0]
+	suite.Equal("image/jpeg", attachment1.File.ContentType)
+	suite.Equal("https://s3-us-west-2.amazonaws.com/plushcity/media_attachments/files/106/867/380/219/163/828/original/88e8758c5f011439.jpg", attachment1.RemoteURL)
+	suite.Equal("It's a cute plushie.", attachment1.Description)
+	suite.Empty(attachment1.Blurhash) // atm we discard blurhashes and generate them ourselves during processing
+}
+
+func (suite *ExtractTestSuite) TestExtractNoAttachments() {
+	note := streams.NewActivityStreamsNote()
+
+	attachments, err := ap.ExtractAttachments(note)
+	suite.NoError(err)
+	suite.Empty(attachments)
+}
+
+func (suite *ExtractTestSuite) TestExtractAttachmentsMissingContentType() {
+	d1 := document1()
+	d1.SetActivityStreamsMediaType(streams.NewActivityStreamsMediaTypeProperty())
+
+	a1 := streams.NewActivityStreamsAttachmentProperty()
+	a1.AppendActivityStreamsDocument(d1)
+
+	note := streams.NewActivityStreamsNote()
+	note.SetActivityStreamsAttachment(a1)
+
+	attachments, err := ap.ExtractAttachments(note)
+	suite.NoError(err)
+	suite.Empty(attachments)
+}
+
+func (suite *ExtractTestSuite) TestExtractAttachmentMissingContentType() {
+
+	d1 := document1()
+	d1.SetActivityStreamsMediaType(streams.NewActivityStreamsMediaTypeProperty())
+
+	attachment, err := ap.ExtractAttachment(d1)
+	suite.EqualError(err, "no media type")
+	suite.Nil(attachment)
+}
+
+func (suite *ExtractTestSuite) TestExtractAttachmentMissingURL() {
+	d1 := document1()
+	d1.SetActivityStreamsUrl(streams.NewActivityStreamsUrlProperty())
+
+	attachment, err := ap.ExtractAttachment(d1)
+	suite.EqualError(err, "could not extract url")
+	suite.Nil(attachment)
+}
+
+func TestExtractTestSuite(t *testing.T) {
+	suite.Run(t, &ExtractTestSuite{})
+}

internal/federation/dereferencing/attachment.go

@@ -28,17 +28,23 @@
 	"github.com/superseriousbusiness/gotosocial/internal/gtsmodel"
 )
 
-func (d *deref) GetRemoteAttachment(ctx context.Context, requestingUsername string, remoteAttachmentURI *url.URL, ownerAccountID string, statusID string, expectedContentType string) (*gtsmodel.MediaAttachment, error) {
+func (d *deref) GetRemoteAttachment(ctx context.Context, requestingUsername string, minAttachment *gtsmodel.MediaAttachment) (*gtsmodel.MediaAttachment, error) {
+	if minAttachment.RemoteURL == "" {
+		return nil, fmt.Errorf("GetRemoteAttachment: minAttachment remote URL was empty")
+	}
+	remoteAttachmentURL := minAttachment.RemoteURL
+
 	l := d.log.WithFields(logrus.Fields{
 		"username":            requestingUsername,
-		"remoteAttachmentURI": remoteAttachmentURI,
+		"remoteAttachmentURL": remoteAttachmentURL,
 	})
 
+	// return early if we already have the attachment somewhere
 	maybeAttachment := &gtsmodel.MediaAttachment{}
 	where := []db.Where{
 		{
 			Key:   "remote_url",
-			Value: remoteAttachmentURI.String(),
+			Value: remoteAttachmentURL,
 		},
 	}
 
@@ -48,12 +54,11 @@ func (d *deref) GetRemoteAttachment(ctx context.Context, requestingUsername stri
 		return maybeAttachment, nil
 	}
 
-	a, err := d.RefreshAttachment(ctx, requestingUsername, remoteAttachmentURI, ownerAccountID, expectedContentType)
+	a, err := d.RefreshAttachment(ctx, requestingUsername, minAttachment)
 	if err != nil {
 		return nil, fmt.Errorf("GetRemoteAttachment: error refreshing attachment: %s", err)
 	}
 
-	a.StatusID = statusID
 	if err := d.db.Put(ctx, a); err != nil {
 		if err != db.ErrAlreadyExists {
 			return nil, fmt.Errorf("GetRemoteAttachment: error inserting attachment: %s", err)
@@ -63,19 +68,32 @@ func (d *deref) GetRemoteAttachment(ctx context.Context, requestingUsername stri
 	return a, nil
 }
 
-func (d *deref) RefreshAttachment(ctx context.Context, requestingUsername string, remoteAttachmentURI *url.URL, ownerAccountID string, expectedContentType string) (*gtsmodel.MediaAttachment, error) {
+func (d *deref) RefreshAttachment(ctx context.Context, requestingUsername string, minAttachment *gtsmodel.MediaAttachment) (*gtsmodel.MediaAttachment, error) {
 	// it just doesn't exist or we have to refresh
+	if minAttachment.AccountID == "" {
+		return nil, fmt.Errorf("RefreshAttachment: minAttachment account ID was empty")
+	}
+
+	if minAttachment.File.ContentType == "" {
+		return nil, fmt.Errorf("RefreshAttachment: minAttachment.file.contentType was empty")
+	}
+
 	t, err := d.transportController.NewTransportForUsername(ctx, requestingUsername)
 	if err != nil {
 		return nil, fmt.Errorf("RefreshAttachment: error creating transport: %s", err)
 	}
 
-	attachmentBytes, err := t.DereferenceMedia(ctx, remoteAttachmentURI, expectedContentType)
+	derefURI, err := url.Parse(minAttachment.RemoteURL)
+	if err != nil {
+		return nil, err
+	}
+
+	attachmentBytes, err := t.DereferenceMedia(ctx, derefURI, minAttachment.File.ContentType)
 	if err != nil {
 		return nil, fmt.Errorf("RefreshAttachment: error dereferencing media: %s", err)
 	}
 
-	a, err := d.mediaHandler.ProcessAttachment(ctx, attachmentBytes, ownerAccountID, remoteAttachmentURI.String())
+	a, err := d.mediaHandler.ProcessAttachment(ctx, attachmentBytes, minAttachment)
 	if err != nil {
 		return nil, fmt.Errorf("RefreshAttachment: error processing attachment: %s", err)
 	}

internal/federation/dereferencing/attachment_test.go

@@ -0,0 +1,106 @@
+/*
+   GoToSocial
+   Copyright (C) 2021 GoToSocial Authors admin@gotosocial.org
+
+   This program is free software: you can redistribute it and/or modify
+   it under the terms of the GNU Affero General Public License as published by
+   the Free Software Foundation, either version 3 of the License, or
+   (at your option) any later version.
+
+   This program is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+   GNU Affero General Public License for more details.
+
+   You should have received a copy of the GNU Affero General Public License
+   along with this program.  If not, see <http://www.gnu.org/licenses/>.
+*/
+
+package dereferencing_test
+
+import (
+	"context"
+	"testing"
+
+	"github.com/stretchr/testify/suite"
+	"github.com/superseriousbusiness/gotosocial/internal/gtsmodel"
+)
+
+type AttachmentTestSuite struct {
+	DereferencerStandardTestSuite
+}
+
+func (suite *AttachmentTestSuite) TestDereferenceAttachmentOK() {
+	fetchingAccount := suite.testAccounts["local_account_1"]
+
+	attachmentOwner := "01FENS9F666SEQ6TYQWEEY78GM"
+	attachmentStatus := "01FENS9NTTVNEX1YZV7GB63MT8"
+	attachmentContentType := "image/jpeg"
+	attachmentURL := "https://s3-us-west-2.amazonaws.com/plushcity/media_attachments/files/106/867/380/219/163/828/original/88e8758c5f011439.jpg"
+	attachmentDescription := "It's a cute plushie."
+
+	minAttachment := &gtsmodel.MediaAttachment{
+		RemoteURL: attachmentURL,
+		AccountID: attachmentOwner,
+		StatusID:  attachmentStatus,
+		File: gtsmodel.File{
+			ContentType: attachmentContentType,
+		},
+		Description: attachmentDescription,
+	}
+
+	attachment, err := suite.dereferencer.GetRemoteAttachment(context.Background(), fetchingAccount.Username, minAttachment)
+	suite.NoError(err)
+	suite.NotNil(attachment)
+
+	suite.Equal(attachmentOwner, attachment.AccountID)
+	suite.Equal(attachmentStatus, attachment.StatusID)
+	suite.Equal(attachmentURL, attachment.RemoteURL)
+	suite.NotEmpty(attachment.URL)
+	suite.NotEmpty(attachment.Blurhash)
+	suite.NotEmpty(attachment.ID)
+	suite.NotEmpty(attachment.CreatedAt)
+	suite.NotEmpty(attachment.UpdatedAt)
+	suite.Equal(1.336546184738956, attachment.FileMeta.Original.Aspect)
+	suite.Equal(2071680, attachment.FileMeta.Original.Size)
+	suite.Equal(1245, attachment.FileMeta.Original.Height)
+	suite.Equal(1664, attachment.FileMeta.Original.Width)
+	suite.Equal("LwQ9yKRP_4t8t7RjWBt7%hozM_ay", attachment.Blurhash)
+	suite.Equal(gtsmodel.ProcessingStatusProcessed, attachment.Processing)
+	suite.NotEmpty(attachment.File.Path)
+	suite.Equal(attachmentContentType, attachment.File.ContentType)
+	suite.Equal(attachmentDescription, attachment.Description)
+
+	suite.NotEmpty(attachment.Thumbnail.Path)
+	suite.NotEmpty(attachment.Type)
+
+	// attachment should also now be in the database
+	dbAttachment, err := suite.db.GetAttachmentByID(context.Background(), attachment.ID)
+	suite.NoError(err)
+	suite.NotNil(dbAttachment)
+
+	suite.Equal(attachmentOwner, dbAttachment.AccountID)
+	suite.Equal(attachmentStatus, dbAttachment.StatusID)
+	suite.Equal(attachmentURL, dbAttachment.RemoteURL)
+	suite.NotEmpty(dbAttachment.URL)
+	suite.NotEmpty(dbAttachment.Blurhash)
+	suite.NotEmpty(dbAttachment.ID)
+	suite.NotEmpty(dbAttachment.CreatedAt)
+	suite.NotEmpty(dbAttachment.UpdatedAt)
+	suite.Equal(1.336546184738956, dbAttachment.FileMeta.Original.Aspect)
+	suite.Equal(2071680, dbAttachment.FileMeta.Original.Size)
+	suite.Equal(1245, dbAttachment.FileMeta.Original.Height)
+	suite.Equal(1664, dbAttachment.FileMeta.Original.Width)
+	suite.Equal("LwQ9yKRP_4t8t7RjWBt7%hozM_ay", dbAttachment.Blurhash)
+	suite.Equal(gtsmodel.ProcessingStatusProcessed, dbAttachment.Processing)
+	suite.NotEmpty(dbAttachment.File.Path)
+	suite.Equal(attachmentContentType, dbAttachment.File.ContentType)
+	suite.Equal(attachmentDescription, dbAttachment.Description)
+
+	suite.NotEmpty(dbAttachment.Thumbnail.Path)
+	suite.NotEmpty(dbAttachment.Type)
+}
+
+func TestAttachmentTestSuite(t *testing.T) {
+	suite.Run(t, new(AttachmentTestSuite))
+}

internal/federation/dereferencing/dereferencer.go

@@ -43,8 +43,34 @@ type Dereferencer interface {
 
 	GetRemoteInstance(ctx context.Context, username string, remoteInstanceURI *url.URL) (*gtsmodel.Instance, error)
 
-	GetRemoteAttachment(ctx context.Context, username string, remoteAttachmentURI *url.URL, ownerAccountID string, statusID string, expectedContentType string) (*gtsmodel.MediaAttachment, error)
-	RefreshAttachment(ctx context.Context, requestingUsername string, remoteAttachmentURI *url.URL, ownerAccountID string, expectedContentType string) (*gtsmodel.MediaAttachment, error)
+	// GetRemoteAttachment takes a minimal attachment struct and converts it into a fully fleshed out attachment, stored in the database and instance storage.
+	//
+	// The parameter minAttachment must have at least the following fields defined:
+	//   * minAttachment.RemoteURL
+	//   * minAttachment.AccountID
+	//   * minAttachment.File.ContentType
+	//
+	// The returned attachment will have an ID generated for it, so no need to generate one beforehand.
+	// A blurhash will also be generated for the attachment.
+	//
+	// Most other fields will be preserved on the passed attachment, including:
+	//   * minAttachment.StatusID
+	//   * minAttachment.CreatedAt
+	//   * minAttachment.UpdatedAt
+	//   * minAttachment.FileMeta
+	//   * minAttachment.AccountID
+	//   * minAttachment.Description
+	//   * minAttachment.ScheduledStatusID
+	//   * minAttachment.Thumbnail.RemoteURL
+	//   * minAttachment.Avatar
+	//   * minAttachment.Header
+	//
+	// GetRemoteAttachment will return early if an attachment with the same value as minAttachment.RemoteURL
+	// is found in the database -- then that attachment will be returned and nothing else will be changed or stored.
+	GetRemoteAttachment(ctx context.Context, requestingUsername string, minAttachment *gtsmodel.MediaAttachment) (*gtsmodel.MediaAttachment, error)
+	// RefreshAttachment is like GetRemoteAttachment, but the attachment will always be dereferenced again,
+	// whether or not it was already stored in the database.
+	RefreshAttachment(ctx context.Context, requestingUsername string, minAttachment *gtsmodel.MediaAttachment) (*gtsmodel.MediaAttachment, error)
 
 	DereferenceAnnounce(ctx context.Context, announce *gtsmodel.Status, requestingUsername string) error
 	DereferenceThread(ctx context.Context, username string, statusIRI *url.URL) error

internal/federation/dereferencing/dereferencer_test.go

@@ -19,24 +19,131 @@
 package dereferencing_test
 
 import (
+	"bytes"
+	"encoding/json"
+	"io"
+	"net/http"
+
+	"github.com/go-fed/activity/streams"
 	"github.com/go-fed/activity/streams/vocab"
 	"github.com/sirupsen/logrus"
 	"github.com/stretchr/testify/suite"
+	"github.com/superseriousbusiness/gotosocial/internal/blob"
 	"github.com/superseriousbusiness/gotosocial/internal/config"
 	"github.com/superseriousbusiness/gotosocial/internal/db"
 	"github.com/superseriousbusiness/gotosocial/internal/federation/dereferencing"
 	"github.com/superseriousbusiness/gotosocial/internal/gtsmodel"
+	"github.com/superseriousbusiness/gotosocial/internal/transport"
+	"github.com/superseriousbusiness/gotosocial/testrig"
 )
 
 type DereferencerStandardTestSuite struct {
 	suite.Suite
-	config *config.Config
-	db     db.DB
-	log    *logrus.Logger
+	config  *config.Config
+	db      db.DB
+	log     *logrus.Logger
+	storage blob.Storage
 
-	testRemoteStatuses map[string]vocab.ActivityStreamsNote
-	testRemoteAccounts map[string]vocab.ActivityStreamsPerson
-	testAccounts       map[string]*gtsmodel.Account
+	testRemoteStatuses    map[string]vocab.ActivityStreamsNote
+	testRemoteAccounts    map[string]vocab.ActivityStreamsPerson
+	testRemoteAttachments map[string]testrig.RemoteAttachmentFile
+	testAccounts          map[string]*gtsmodel.Account
 
 	dereferencer dereferencing.Dereferencer
 }
+
+func (suite *DereferencerStandardTestSuite) SetupSuite() {
+	suite.testAccounts = testrig.NewTestAccounts()
+	suite.testRemoteStatuses = testrig.NewTestFediStatuses()
+	suite.testRemoteAccounts = testrig.NewTestFediPeople()
+	suite.testRemoteAttachments = testrig.NewTestFediAttachments("../../../testrig/media")
+}
+
+func (suite *DereferencerStandardTestSuite) SetupTest() {
+	suite.config = testrig.NewTestConfig()
+	suite.db = testrig.NewTestDB()
+	suite.log = testrig.NewTestLog()
+	suite.storage = testrig.NewTestStorage()
+	suite.dereferencer = dereferencing.NewDereferencer(suite.config,
+		suite.db,
+		testrig.NewTestTypeConverter(suite.db),
+		suite.mockTransportController(),
+		testrig.NewTestMediaHandler(suite.db, suite.storage),
+		suite.log)
+	testrig.StandardDBSetup(suite.db, nil)
+}
+
+func (suite *DereferencerStandardTestSuite) TearDownTest() {
+	testrig.StandardDBTeardown(suite.db)
+}
+
+// mockTransportController returns basically a miniature muxer, which returns a different
+// value based on the request URL. It can be used to return remote statuses, profiles, etc,
+// as though they were actually being dereferenced. If the URL doesn't correspond to any person
+// or note or attachment that we have stored, then just a 200 code will be returned, with an empty body.
+func (suite *DereferencerStandardTestSuite) mockTransportController() transport.Controller {
+	do := func(req *http.Request) (*http.Response, error) {
+		suite.log.Debugf("received request for %s", req.URL)
+
+		responseBytes := []byte{}
+		responseType := ""
+		responseLength := 0
+
+		note, ok := suite.testRemoteStatuses[req.URL.String()]
+		if ok {
+			// the request is for a note that we have stored
+			noteI, err := streams.Serialize(note)
+			if err != nil {
+				panic(err)
+			}
+			noteJson, err := json.Marshal(noteI)
+			if err != nil {
+				panic(err)
+			}
+			responseBytes = noteJson
+			responseType = "application/activity+json"
+		}
+
+		person, ok := suite.testRemoteAccounts[req.URL.String()]
+		if ok {
+			// the request is for a person that we have stored
+			personI, err := streams.Serialize(person)
+			if err != nil {
+				panic(err)
+			}
+			personJson, err := json.Marshal(personI)
+			if err != nil {
+				panic(err)
+			}
+			responseBytes = personJson
+			responseType = "application/activity+json"
+		}
+
+		attachment, ok := suite.testRemoteAttachments[req.URL.String()]
+		if ok {
+			responseBytes = attachment.Data
+			responseType = attachment.ContentType
+		}
+
+		if len(responseBytes) != 0 {
+			// we found something, so print what we're going to return
+			suite.log.Debugf("returning response %s", string(responseBytes))
+		}
+		responseLength = len(responseBytes)
+
+		reader := bytes.NewReader(responseBytes)
+		readCloser := io.NopCloser(reader)
+		response := &http.Response{
+			StatusCode:    200,
+			Body:          readCloser,
+			ContentLength: int64(responseLength),
+			Header: http.Header{
+				"content-type": {responseType},
+			},
+		}
+
+		return response, nil
+	}
+	mockClient := testrig.NewMockHTTPClient(do)
+	return testrig.NewTestTransportController(mockClient, suite.db)
+}

internal/federation/dereferencing/status.go

@@ -396,13 +396,10 @@ func (d *deref) populateStatusAttachments(ctx context.Context, status *gtsmodel.
 	attachments := []*gtsmodel.MediaAttachment{}
 
 	for _, a := range status.Attachments {
-		aURL, err := url.Parse(a.RemoteURL)
-		if err != nil {
-			l.Errorf("populateStatusAttachments: couldn't parse attachment url %s: %s", a.RemoteURL, err)
-			continue
-		}
+		a.AccountID = status.AccountID
+		a.StatusID = status.ID
 
-		attachment, err := d.GetRemoteAttachment(ctx, requestingUsername, aURL, status.AccountID, status.ID, a.File.ContentType)
+		attachment, err := d.GetRemoteAttachment(ctx, requestingUsername, a)
 		if err != nil {
 			l.Errorf("populateStatusAttachments: couldn't get remote attachment %s: %s", a.RemoteURL, err)
 			continue

internal/federation/dereferencing/status_test.go

@@ -19,21 +19,14 @@
 package dereferencing_test
 
 import (
-	"bytes"
 	"context"
-	"encoding/json"
-	"io"
-	"net/http"
 	"testing"
 	"time"
 
-	"github.com/go-fed/activity/streams"
 	"github.com/stretchr/testify/suite"
 	"github.com/superseriousbusiness/gotosocial/internal/ap"
 	"github.com/superseriousbusiness/gotosocial/internal/db"
-	"github.com/superseriousbusiness/gotosocial/internal/federation/dereferencing"
 	"github.com/superseriousbusiness/gotosocial/internal/gtsmodel"
-	"github.com/superseriousbusiness/gotosocial/internal/transport"
 	"github.com/superseriousbusiness/gotosocial/testrig"
 )
 
@@ -41,81 +34,6 @@ type StatusTestSuite struct {
 	DereferencerStandardTestSuite
 }
 
-// mockTransportController returns basically a miniature muxer, which returns a different
-// value based on the request URL. It can be used to return remote statuses, profiles, etc,
-// as though they were actually being dereferenced. If the URL doesn't correspond to any person
-// or note or attachment that we have stored, then just a 200 code will be returned, with an empty body.
-func (suite *StatusTestSuite) mockTransportController() transport.Controller {
-	do := func(req *http.Request) (*http.Response, error) {
-		suite.log.Debugf("received request for %s", req.URL)
-
-		responseBytes := []byte{}
-
-		note, ok := suite.testRemoteStatuses[req.URL.String()]
-		if ok {
-			// the request is for a note that we have stored
-			noteI, err := streams.Serialize(note)
-			if err != nil {
-				panic(err)
-			}
-			noteJson, err := json.Marshal(noteI)
-			if err != nil {
-				panic(err)
-			}
-			responseBytes = noteJson
-		}
-
-		person, ok := suite.testRemoteAccounts[req.URL.String()]
-		if ok {
-			// the request is for a person that we have stored
-			personI, err := streams.Serialize(person)
-			if err != nil {
-				panic(err)
-			}
-			personJson, err := json.Marshal(personI)
-			if err != nil {
-				panic(err)
-			}
-			responseBytes = personJson
-		}
-
-		if len(responseBytes) != 0 {
-			// we found something, so print what we're going to return
-			suite.log.Debugf("returning response %s", string(responseBytes))
-		}
-
-		reader := bytes.NewReader(responseBytes)
-		readCloser := io.NopCloser(reader)
-		response := &http.Response{
-			StatusCode: 200,
-			Body:       readCloser,
-		}
-
-		return response, nil
-	}
-	mockClient := testrig.NewMockHTTPClient(do)
-	return testrig.NewTestTransportController(mockClient, suite.db)
-}
-
-func (suite *StatusTestSuite) SetupSuite() {
-	suite.testAccounts = testrig.NewTestAccounts()
-	suite.testRemoteStatuses = testrig.NewTestFediStatuses()
-	suite.testRemoteAccounts = testrig.NewTestFediPeople()
-}
-
-func (suite *StatusTestSuite) SetupTest() {
-	suite.config = testrig.NewTestConfig()
-	suite.db = testrig.NewTestDB()
-	suite.log = testrig.NewTestLog()
-	suite.dereferencer = dereferencing.NewDereferencer(suite.config,
-		suite.db,
-		testrig.NewTestTypeConverter(suite.db),
-		suite.mockTransportController(),
-		testrig.NewTestMediaHandler(suite.db, testrig.NewTestStorage()),
-		suite.log)
-	testrig.StandardDBSetup(suite.db, nil)
-}
-
 func (suite *StatusTestSuite) TestDereferenceSimpleStatus() {
 	fetchingAccount := suite.testAccounts["local_account_1"]
 
@@ -205,10 +123,6 @@ func (suite *StatusTestSuite) TestDereferenceStatusWithMention() {
 	suite.False(m.Silent)
 }
 
-func (suite *StatusTestSuite) TearDownTest() {
-	testrig.StandardDBTeardown(suite.db)
-}
-
 func TestStatusTestSuite(t *testing.T) {
 	suite.Run(t, new(StatusTestSuite))
 }

internal/media/handler.go

@@ -73,7 +73,7 @@ type Handler interface {
 	// puts it in whatever storage backend we're using, sets the relevant fields in the database for the new media,
 	// and then returns information to the caller about the attachment. It's the caller's responsibility to put the returned struct
 	// in the database.
-	ProcessAttachment(ctx context.Context, attachment []byte, accountID string, remoteURL string) (*gtsmodel.MediaAttachment, error)
+	ProcessAttachment(ctx context.Context, attachmentBytes []byte, minAttachment *gtsmodel.MediaAttachment) (*gtsmodel.MediaAttachment, error)
 
 	// ProcessLocalEmoji takes a new emoji and a shortcode, cleans it up, puts it in storage, and creates a new
 	// *gts.Emoji for it, then returns it to the caller. It's the caller's responsibility to put the returned struct
@@ -145,11 +145,14 @@ func (mh *mediaHandler) ProcessHeaderOrAvatar(ctx context.Context, attachment []
 // ProcessAttachment takes a new attachment and the owning account, checks it out, removes exif data from it,
 // puts it in whatever storage backend we're using, sets the relevant fields in the database for the new media,
 // and then returns information to the caller about the attachment.
-func (mh *mediaHandler) ProcessAttachment(ctx context.Context, attachment []byte, accountID string, remoteURL string) (*gtsmodel.MediaAttachment, error) {
-	contentType, err := parseContentType(attachment)
+func (mh *mediaHandler) ProcessAttachment(ctx context.Context, attachmentBytes []byte, minAttachment *gtsmodel.MediaAttachment) (*gtsmodel.MediaAttachment, error) {
+	contentType, err := parseContentType(attachmentBytes)
 	if err != nil {
 		return nil, err
 	}
+
+	minAttachment.File.ContentType = contentType
+
 	mainType := strings.Split(contentType, "/")[0]
 	switch mainType {
 	// case MIMEVideo:
@@ -164,10 +167,10 @@ func (mh *mediaHandler) ProcessAttachment(ctx context.Context, attachment []byte
 		if !SupportedImageType(contentType) {
 			return nil, fmt.Errorf("image type %s not supported", contentType)
 		}
-		if len(attachment) == 0 {
+		if len(attachmentBytes) == 0 {
 			return nil, errors.New("image was of size 0")
 		}
-		return mh.processImageAttachment(attachment, accountID, contentType, remoteURL)
+		return mh.processImageAttachment(attachmentBytes, minAttachment)
 	default:
 		break
 	}

internal/media/processimage.go

@@ -28,12 +28,14 @@
 	"github.com/superseriousbusiness/gotosocial/internal/id"
 )
 
-func (mh *mediaHandler) processImageAttachment(data []byte, accountID string, contentType string, remoteURL string) (*gtsmodel.MediaAttachment, error) {
+func (mh *mediaHandler) processImageAttachment(data []byte, minAttachment *gtsmodel.MediaAttachment) (*gtsmodel.MediaAttachment, error) {
 	var clean []byte
 	var err error
 	var original *imageAndMeta
 	var small *imageAndMeta
 
+	contentType := minAttachment.File.ContentType
+
 	switch contentType {
 	case MIMEJpeg, MIMEPng:
 		if clean, err = purgeExif(data); err != nil {
@@ -66,46 +68,47 @@ func (mh *mediaHandler) processImageAttachment(data []byte, accountID string, co
 	}
 
 	URLbase := fmt.Sprintf("%s://%s%s", mh.config.StorageConfig.ServeProtocol, mh.config.StorageConfig.ServeHost, mh.config.StorageConfig.ServeBasePath)
-	originalURL := fmt.Sprintf("%s/%s/attachment/original/%s.%s", URLbase, accountID, newMediaID, extension)
-	smallURL := fmt.Sprintf("%s/%s/attachment/small/%s.jpeg", URLbase, accountID, newMediaID) // all thumbnails/smalls are encoded as jpeg
+	originalURL := fmt.Sprintf("%s/%s/attachment/original/%s.%s", URLbase, minAttachment.AccountID, newMediaID, extension)
+	smallURL := fmt.Sprintf("%s/%s/attachment/small/%s.jpeg", URLbase, minAttachment.AccountID, newMediaID) // all thumbnails/smalls are encoded as jpeg
 
 	// we store the original...
-	originalPath := fmt.Sprintf("%s/%s/%s/%s/%s.%s", mh.config.StorageConfig.BasePath, accountID, Attachment, Original, newMediaID, extension)
+	originalPath := fmt.Sprintf("%s/%s/%s/%s/%s.%s", mh.config.StorageConfig.BasePath, minAttachment.AccountID, Attachment, Original, newMediaID, extension)
 	if err := mh.storage.StoreFileAt(originalPath, original.image); err != nil {
 		return nil, fmt.Errorf("storage error: %s", err)
 	}
 
 	// and a thumbnail...
-	smallPath := fmt.Sprintf("%s/%s/%s/%s/%s.jpeg", mh.config.StorageConfig.BasePath, accountID, Attachment, Small, newMediaID) // all thumbnails/smalls are encoded as jpeg
+	smallPath := fmt.Sprintf("%s/%s/%s/%s/%s.jpeg", mh.config.StorageConfig.BasePath, minAttachment.AccountID, Attachment, Small, newMediaID) // all thumbnails/smalls are encoded as jpeg
 	if err := mh.storage.StoreFileAt(smallPath, small.image); err != nil {
 		return nil, fmt.Errorf("storage error: %s", err)
 	}
 
-	ma := &gtsmodel.MediaAttachment{
-		ID:        newMediaID,
-		StatusID:  "",
-		URL:       originalURL,
-		RemoteURL: remoteURL,
-		CreatedAt: time.Now(),
-		UpdatedAt: time.Now(),
-		Type:      gtsmodel.FileTypeImage,
-		FileMeta: gtsmodel.FileMeta{
-			Original: gtsmodel.Original{
-				Width:  original.width,
-				Height: original.height,
-				Size:   original.size,
-				Aspect: original.aspect,
-			},
-			Small: gtsmodel.Small{
-				Width:  small.width,
-				Height: small.height,
-				Size:   small.size,
-				Aspect: small.aspect,
-			},
-		},
-		AccountID:         accountID,
-		Description:       "",
-		ScheduledStatusID: "",
+	minAttachment.FileMeta.Original = gtsmodel.Original{
+		Width:  original.width,
+		Height: original.height,
+		Size:   original.size,
+		Aspect: original.aspect,
+	}
+
+	minAttachment.FileMeta.Small = gtsmodel.Small{
+		Width:  small.width,
+		Height: small.height,
+		Size:   small.size,
+		Aspect: small.aspect,
+	}
+
+	attachment := &gtsmodel.MediaAttachment{
+		ID:                newMediaID,
+		StatusID:          minAttachment.StatusID,
+		URL:               originalURL,
+		RemoteURL:         minAttachment.RemoteURL,
+		CreatedAt:         minAttachment.CreatedAt,
+		UpdatedAt:         minAttachment.UpdatedAt,
+		Type:              gtsmodel.FileTypeImage,
+		FileMeta:          minAttachment.FileMeta,
+		AccountID:         minAttachment.AccountID,
+		Description:       minAttachment.Description,
+		ScheduledStatusID: minAttachment.ScheduledStatusID,
 		Blurhash:          original.blurhash,
 		Processing:        2,
 		File: gtsmodel.File{
@@ -120,12 +123,12 @@ func (mh *mediaHandler) processImageAttachment(data []byte, accountID string, co
 			FileSize:    len(small.image),
 			UpdatedAt:   time.Now(),
 			URL:         smallURL,
-			RemoteURL:   "",
+			RemoteURL:   minAttachment.Thumbnail.RemoteURL,
 		},
-		Avatar: false,
-		Header: false,
+		Avatar: minAttachment.Avatar,
+		Header: minAttachment.Header,
 	}
 
-	return ma, nil
+	return attachment, nil
 
 }

internal/processing/media/create.go

@@ -24,6 +24,7 @@
 	"errors"
 	"fmt"
 	"io"
+	"time"
 
 	apimodel "github.com/superseriousbusiness/gotosocial/internal/api/model"
 	"github.com/superseriousbusiness/gotosocial/internal/gtsmodel"
@@ -45,25 +46,30 @@ func (p *processor) Create(ctx context.Context, account *gtsmodel.Account, form
 		return nil, errors.New("could not read provided attachment: size 0 bytes")
 	}
 
-	// allow the mediaHandler to work its magic of processing the attachment bytes, and putting them in whatever storage backend we're using
-	attachment, err := p.mediaHandler.ProcessAttachment(ctx, buf.Bytes(), account.ID, "")
-	if err != nil {
-		return nil, fmt.Errorf("error reading attachment: %s", err)
-	}
-
-	// now we need to add extra fields that the attachment processor doesn't know (from the form)
-	// TODO: handle this inside mediaHandler.ProcessAttachment (just pass more params to it)
-
-	// first description
-	attachment.Description = text.RemoveHTML(form.Description) // remove any HTML from the image description
-
 	// now parse the focus parameter
 	focusx, focusy, err := parseFocus(form.Focus)
 	if err != nil {
-		return nil, err
+		return nil, fmt.Errorf("couldn't parse attachment focus: %s", err)
+	}
+
+	minAttachment := &gtsmodel.MediaAttachment{
+		CreatedAt:   time.Now(),
+		UpdatedAt:   time.Now(),
+		AccountID:   account.ID,
+		Description: text.RemoveHTML(form.Description),
+		FileMeta: gtsmodel.FileMeta{
+			Focus: gtsmodel.Focus{
+				X: focusx,
+				Y: focusy,
+			},
+		},
+	}
+
+	// allow the mediaHandler to work its magic of processing the attachment bytes, and putting them in whatever storage backend we're using
+	attachment, err := p.mediaHandler.ProcessAttachment(ctx, buf.Bytes(), minAttachment)
+	if err != nil {
+		return nil, fmt.Errorf("error reading attachment: %s", err)
 	}
-	attachment.FileMeta.Focus.X = focusx
-	attachment.FileMeta.Focus.Y = focusy
 
 	// prepare the frontend representation now -- if there are any errors here at least we can bail without
 	// having already put something in the database and then having to clean it up again (eugh)

testrig/testmodels.go

@@ -27,10 +27,12 @@
 	"crypto/x509"
 	"encoding/json"
 	"encoding/pem"
+	"fmt"
 	"io/ioutil"
 	"net"
 	"net/http"
 	"net/url"
+	"os"
 	"time"
 
 	"github.com/go-fed/activity/pub"
@@ -1285,6 +1287,25 @@ func NewTestFediPeople() map[string]vocab.ActivityStreamsPerson {
 	}
 }
 
+// RemoteAttachmentFile mimics a remote (federated) attachment
+type RemoteAttachmentFile struct {
+	Data        []byte
+	ContentType string
+}
+
+func NewTestFediAttachments(relativePath string) map[string]RemoteAttachmentFile {
+	beeBytes, err := os.ReadFile(fmt.Sprintf("%s/beeplushie.jpg", relativePath))
+	if err != nil {
+		panic(err)
+	}
+	return map[string]RemoteAttachmentFile{
+		"https://s3-us-west-2.amazonaws.com/plushcity/media_attachments/files/106/867/380/219/163/828/original/88e8758c5f011439.jpg": {
+			Data:        beeBytes,
+			ContentType: "image/jpeg",
+		},
+	}
+}
+
 func NewTestFediStatuses() map[string]vocab.ActivityStreamsNote {
 	return map[string]vocab.ActivityStreamsNote{
 		"https://unknown-instance.com/users/brand_new_person/statuses/01FE4NTHKWW7THT67EF10EB839": newNote(