[feature] Clean up/uncache remote media (#407)

* Add whereNotEmptyAndNotNull * Add GetRemoteOlderThanDays * Add GetRemoteOlderThanDays * Add PruneRemote to Manager interface * Start implementing PruneRemote * add new attachment + status to tests * fix up and test GetRemoteOlderThan * fix bad import * PruneRemote: return number pruned * add Cached column to mediaattachment * update + test pruneRemote * update mediaTest * use Cached column * upstep bun to latest version * embed structs in mediaAttachment * migrate mediaAttachment to new format * don't default cached to true * select only remote media * update db dependencies * step bun back to last working version * update pruneRemote to use Cached field * fix storage path of test attachments * add recache logic to manager * fix trimmed aspect ratio * test prune and recache * return errwithcode * tidy up different paths for emoji vs attachment * fix incorrect thumbnail type being stored * expose TransportController to media processor * implement tee-ing recached content * add thoughts of dog to test fedi attachments * test get remote files * add comment on PruneRemote * add postData cleanup to recache * test thumbnail fetching * add incredible diagram * go mod tidy * buffer pipes for recache streaming * test for client stops reading after 1kb * add media-remote-cache-days to config * add cron package * wrap logrus so it's available to cron * start and stop cron jobs gracefully
2025-06-05 21:59:39 +02:00 · 2022-03-07 11:08:26 +01:00
parent 100f1280a6
commit 07727753b9
424 changed files with 637100 additions and 176498 deletions
--- a/internal/processing/media.go
+++ b/internal/processing/media.go
@@ -38,6 +38,6 @@ func (p *processor) MediaUpdate(ctx context.Context, authed *oauth.Auth, mediaAt
 	return p.mediaProcessor.Update(ctx, authed.Account, mediaAttachmentID, form)
 }

-func (p *processor) FileGet(ctx context.Context, authed *oauth.Auth, form *apimodel.GetContentRequestForm) (*apimodel.Content, error) {
+func (p *processor) FileGet(ctx context.Context, authed *oauth.Auth, form *apimodel.GetContentRequestForm) (*apimodel.Content, gtserror.WithCode) {
 	return p.mediaProcessor.GetFile(ctx, authed.Account, form)
 }
--- a/internal/processing/media/getfile.go
+++ b/internal/processing/media/getfile.go
@@ -19,8 +19,11 @@
 package media

 import (
+	"bufio"
 	"context"
 	"fmt"
+	"io"
+	"net/url"
 	"strings"

 	apimodel "github.com/superseriousbusiness/gotosocial/internal/api/model"
@@ -29,7 +32,7 @@ import (
 	"github.com/superseriousbusiness/gotosocial/internal/media"
 )

-func (p *processor) GetFile(ctx context.Context, account *gtsmodel.Account, form *apimodel.GetContentRequestForm) (*apimodel.Content, error) {
+func (p *processor) GetFile(ctx context.Context, account *gtsmodel.Account, form *apimodel.GetContentRequestForm) (*apimodel.Content, gtserror.WithCode) {
 	// parse the form fields
 	mediaSize, err := media.ParseMediaSize(form.MediaSize)
 	if err != nil {
@@ -46,74 +49,208 @@ func (p *processor) GetFile(ctx context.Context, account *gtsmodel.Account, form
 		return nil, gtserror.NewErrorNotFound(fmt.Errorf("file name %s not parseable", form.FileName))
 	}
 	wantedMediaID := spl[0]
+	expectedAccountID := form.AccountID

 	// get the account that owns the media and make sure it's not suspended
-	acct, err := p.db.GetAccountByID(ctx, form.AccountID)
+	acct, err := p.db.GetAccountByID(ctx, expectedAccountID)
 	if err != nil {
-		return nil, gtserror.NewErrorNotFound(fmt.Errorf("account with id %s could not be selected from the db: %s", form.AccountID, err))
+		return nil, gtserror.NewErrorNotFound(fmt.Errorf("account with id %s could not be selected from the db: %s", expectedAccountID, err))
 	}
 	if !acct.SuspendedAt.IsZero() {
-		return nil, gtserror.NewErrorNotFound(fmt.Errorf("account with id %s is suspended", form.AccountID))
+		return nil, gtserror.NewErrorNotFound(fmt.Errorf("account with id %s is suspended", expectedAccountID))
 	}

 	// make sure the requesting account and the media account don't block each other
 	if account != nil {
-		blocked, err := p.db.IsBlocked(ctx, account.ID, form.AccountID, true)
+		blocked, err := p.db.IsBlocked(ctx, account.ID, expectedAccountID, true)
 		if err != nil {
-			return nil, gtserror.NewErrorNotFound(fmt.Errorf("block status could not be established between accounts %s and %s: %s", form.AccountID, account.ID, err))
+			return nil, gtserror.NewErrorNotFound(fmt.Errorf("block status could not be established between accounts %s and %s: %s", expectedAccountID, account.ID, err))
 		}
 		if blocked {
-			return nil, gtserror.NewErrorNotFound(fmt.Errorf("block exists between accounts %s and %s", form.AccountID, account.ID))
+			return nil, gtserror.NewErrorNotFound(fmt.Errorf("block exists between accounts %s and %s", expectedAccountID, account.ID))
 		}
 	}

 	// the way we store emojis is a little different from the way we store other attachments,
 	// so we need to take different steps depending on the media type being requested
-	content := &apimodel.Content{}
-	var storagePath string
 	switch mediaType {
 	case media.TypeEmoji:
-		e := &gtsmodel.Emoji{}
-		if err := p.db.GetByID(ctx, wantedMediaID, e); err != nil {
-			return nil, gtserror.NewErrorNotFound(fmt.Errorf("emoji %s could not be taken from the db: %s", wantedMediaID, err))
-		}
-		if e.Disabled {
-			return nil, gtserror.NewErrorNotFound(fmt.Errorf("emoji %s has been disabled", wantedMediaID))
-		}
-		switch mediaSize {
-		case media.SizeOriginal:
-			content.ContentType = e.ImageContentType
-			content.ContentLength = int64(e.ImageFileSize)
-			storagePath = e.ImagePath
-		case media.SizeStatic:
-			content.ContentType = e.ImageStaticContentType
-			content.ContentLength = int64(e.ImageStaticFileSize)
-			storagePath = e.ImageStaticPath
-		default:
-			return nil, gtserror.NewErrorNotFound(fmt.Errorf("media size %s not recognized for emoji", mediaSize))
-		}
+		return p.getEmojiContent(ctx, wantedMediaID, mediaSize)
 	case media.TypeAttachment, media.TypeHeader, media.TypeAvatar:
-		a, err := p.db.GetAttachmentByID(ctx, wantedMediaID)
-		if err != nil {
-			return nil, gtserror.NewErrorNotFound(fmt.Errorf("attachment %s could not be taken from the db: %s", wantedMediaID, err))
+		return p.getAttachmentContent(ctx, account, wantedMediaID, expectedAccountID, mediaSize)
+	default:
+		return nil, gtserror.NewErrorNotFound(fmt.Errorf("media type %s not recognized", mediaType))
+	}
+}
+
+func (p *processor) getAttachmentContent(ctx context.Context, requestingAccount *gtsmodel.Account, wantedMediaID string, expectedAccountID string, mediaSize media.Size) (*apimodel.Content, gtserror.WithCode) {
+	attachmentContent := &apimodel.Content{}
+	var storagePath string
+
+	// retrieve attachment from the database and do basic checks on it
+	a, err := p.db.GetAttachmentByID(ctx, wantedMediaID)
+	if err != nil {
+		return nil, gtserror.NewErrorNotFound(fmt.Errorf("attachment %s could not be taken from the db: %s", wantedMediaID, err))
+	}
+
+	if a.AccountID != expectedAccountID {
+		return nil, gtserror.NewErrorNotFound(fmt.Errorf("attachment %s is not owned by %s", wantedMediaID, expectedAccountID))
+	}
+
+	// get file information from the attachment depending on the requested media size
+	switch mediaSize {
+	case media.SizeOriginal:
+		attachmentContent.ContentType = a.File.ContentType
+		attachmentContent.ContentLength = int64(a.File.FileSize)
+		storagePath = a.File.Path
+	case media.SizeSmall:
+		attachmentContent.ContentType = a.Thumbnail.ContentType
+		attachmentContent.ContentLength = int64(a.Thumbnail.FileSize)
+		storagePath = a.Thumbnail.Path
+	default:
+		return nil, gtserror.NewErrorNotFound(fmt.Errorf("media size %s not recognized for attachment", mediaSize))
+	}
+
+	// if we have the media cached on our server already, we can now simply return it from storage
+	if a.Cached {
+		return p.streamFromStorage(storagePath, attachmentContent)
+	}
+
+	// if we don't have it cached, then we can assume two things:
+	// 1. this is remote media, since local media should never be uncached
+	// 2. we need to fetch it again using a transport and the media manager
+	remoteMediaIRI, err := url.Parse(a.RemoteURL)
+	if err != nil {
+		return nil, gtserror.NewErrorNotFound(fmt.Errorf("error parsing remote media iri %s: %s", a.RemoteURL, err))
+	}
+
+	// use an empty string as requestingUsername to use the instance account, unless the request for this
+	// media has been http signed, then use the requesting account to make the request to remote server
+	var requestingUsername string
+	if requestingAccount != nil {
+		requestingUsername = requestingAccount.Username
+	}
+
+	var data media.DataFunc
+	var postDataCallback media.PostDataCallbackFunc
+
+	if mediaSize == media.SizeSmall {
+		// if it's the thumbnail that's requested then the user will have to wait a bit while we process the
+		// large version and derive a thumbnail from it, so use the normal recaching procedure: fetch the media,
+		// process it, then return the thumbnail data
+		data = func(innerCtx context.Context) (io.Reader, int, error) {
+			transport, err := p.transportController.NewTransportForUsername(innerCtx, requestingUsername)
+			if err != nil {
+				return nil, 0, err
+			}
+			return transport.DereferenceMedia(innerCtx, remoteMediaIRI)
 		}
-		if a.AccountID != form.AccountID {
-			return nil, gtserror.NewErrorNotFound(fmt.Errorf("attachment %s is not owned by %s", wantedMediaID, form.AccountID))
+	} else {
+		// if it's the full-sized version being requested, we can cheat a bit by streaming data to the user as
+		// it's retrieved from the remote server, using tee; this saves the user from having to wait while
+		// we process the media on our side
+		//
+		// this looks a bit like this:
+		//
+		//                http fetch                   buffered pipe
+		// remote server ------------> data function ----------------> api caller
+		//                                   |
+		//                                   | tee
+		//                                   |
+		//                                   ▼
+		//                            instance storage
+
+		// Buffer each end of the pipe, so that if the caller drops the connection during the flow, the tee
+		// reader can continue without having to worry about tee-ing into a closed or blocked pipe.
+		pipeReader, pipeWriter := io.Pipe()
+		bufferedWriter := bufio.NewWriterSize(pipeWriter, int(attachmentContent.ContentLength))
+		bufferedReader := bufio.NewReaderSize(pipeReader, int(attachmentContent.ContentLength))
+
+		// the caller will read from the buffered reader, so it doesn't matter if they drop out without reading everything
+		attachmentContent.Content = bufferedReader
+
+		data = func(innerCtx context.Context) (io.Reader, int, error) {
+			transport, err := p.transportController.NewTransportForUsername(innerCtx, requestingUsername)
+			if err != nil {
+				return nil, 0, err
+			}
+
+			readCloser, fileSize, err := transport.DereferenceMedia(innerCtx, remoteMediaIRI)
+			if err != nil {
+				return nil, 0, err
+			}
+
+			// everything read from the readCloser by the media manager will be written into the bufferedWriter
+			teeReader := io.TeeReader(readCloser, bufferedWriter)
+			return teeReader, fileSize, nil
 		}
-		switch mediaSize {
-		case media.SizeOriginal:
-			content.ContentType = a.File.ContentType
-			content.ContentLength = int64(a.File.FileSize)
-			storagePath = a.File.Path
-		case media.SizeSmall:
-			content.ContentType = a.Thumbnail.ContentType
-			content.ContentLength = int64(a.Thumbnail.FileSize)
-			storagePath = a.Thumbnail.Path
-		default:
-			return nil, gtserror.NewErrorNotFound(fmt.Errorf("media size %s not recognized for attachment", mediaSize))
+
+		// close the pipewriter after data has been piped into it, so the reader on the other side doesn't block;
+		// we don't need to close the reader here because that's the caller's responsibility
+		postDataCallback = func(innerCtx context.Context) error {
+			// flush the buffered writer into the buffer of the reader...
+			if err := bufferedWriter.Flush(); err != nil {
+				return err
+			}
+
+			// and close the underlying pipe writer
+			if err := pipeWriter.Close(); err != nil {
+				return err
+			}
+
+			return nil
 		}
 	}

+	// put the media recached in the queue
+	processingMedia, err := p.mediaManager.RecacheMedia(ctx, data, postDataCallback, wantedMediaID)
+	if err != nil {
+		return nil, gtserror.NewErrorNotFound(fmt.Errorf("error recaching media: %s", err))
+	}
+
+	// if it's the thumbnail, stream the processed thumbnail from storage, after waiting for processing to finish
+	if mediaSize == media.SizeSmall {
+		// below function call blocks until all processing on the attachment has finished...
+		if _, err := processingMedia.LoadAttachment(ctx); err != nil {
+			return nil, gtserror.NewErrorNotFound(fmt.Errorf("error loading recached attachment: %s", err))
+		}
+		// ... so now we can safely return it
+		return p.streamFromStorage(storagePath, attachmentContent)
+	}
+
+	return attachmentContent, nil
+}
+
+func (p *processor) getEmojiContent(ctx context.Context, wantedEmojiID string, emojiSize media.Size) (*apimodel.Content, gtserror.WithCode) {
+	emojiContent := &apimodel.Content{}
+	var storagePath string
+
+	e := &gtsmodel.Emoji{}
+	if err := p.db.GetByID(ctx, wantedEmojiID, e); err != nil {
+		return nil, gtserror.NewErrorNotFound(fmt.Errorf("emoji %s could not be taken from the db: %s", wantedEmojiID, err))
+	}
+
+	if e.Disabled {
+		return nil, gtserror.NewErrorNotFound(fmt.Errorf("emoji %s has been disabled", wantedEmojiID))
+	}
+
+	switch emojiSize {
+	case media.SizeOriginal:
+		emojiContent.ContentType = e.ImageContentType
+		emojiContent.ContentLength = int64(e.ImageFileSize)
+		storagePath = e.ImagePath
+	case media.SizeStatic:
+		emojiContent.ContentType = e.ImageStaticContentType
+		emojiContent.ContentLength = int64(e.ImageStaticFileSize)
+		storagePath = e.ImageStaticPath
+	default:
+		return nil, gtserror.NewErrorNotFound(fmt.Errorf("media size %s not recognized for emoji", emojiSize))
+	}
+
+	return p.streamFromStorage(storagePath, emojiContent)
+}
+
+func (p *processor) streamFromStorage(storagePath string, content *apimodel.Content) (*apimodel.Content, gtserror.WithCode) {
 	reader, err := p.storage.GetStream(storagePath)
 	if err != nil {
 		return nil, gtserror.NewErrorNotFound(fmt.Errorf("error retrieving from storage: %s", err))
--- a/internal/processing/media/getfile_test.go
+++ b/internal/processing/media/getfile_test.go
@@ -0,0 +1,208 @@
+/*
+   GoToSocial
+   Copyright (C) 2021-2022 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 media_test
+
+import (
+	"context"
+	"io"
+	"path"
+	"testing"
+	"time"
+
+	"github.com/stretchr/testify/suite"
+	apimodel "github.com/superseriousbusiness/gotosocial/internal/api/model"
+	"github.com/superseriousbusiness/gotosocial/internal/media"
+)
+
+type GetFileTestSuite struct {
+	MediaStandardTestSuite
+}
+
+func (suite *GetFileTestSuite) TestGetRemoteFileCached() {
+	ctx := context.Background()
+
+	testAttachment := suite.testAttachments["remote_account_1_status_1_attachment_1"]
+	fileName := path.Base(testAttachment.File.Path)
+	requestingAccount := suite.testAccounts["local_account_1"]
+
+	content, errWithCode := suite.mediaProcessor.GetFile(ctx, requestingAccount, &apimodel.GetContentRequestForm{
+		AccountID: testAttachment.AccountID,
+		MediaType: string(media.TypeAttachment),
+		MediaSize: string(media.SizeOriginal),
+		FileName:  fileName,
+	})
+
+	suite.NoError(errWithCode)
+	suite.NotNil(content)
+	b, err := io.ReadAll(content.Content)
+	suite.NoError(err)
+
+	if closer, ok := content.Content.(io.Closer); ok {
+		suite.NoError(closer.Close())
+	}
+
+	suite.Equal(suite.testRemoteAttachments[testAttachment.RemoteURL].Data, b)
+	suite.Equal(suite.testRemoteAttachments[testAttachment.RemoteURL].ContentType, content.ContentType)
+	suite.EqualValues(len(suite.testRemoteAttachments[testAttachment.RemoteURL].Data), content.ContentLength)
+}
+
+func (suite *GetFileTestSuite) TestGetRemoteFileUncached() {
+	ctx := context.Background()
+
+	// uncache the file from local
+	testAttachment := suite.testAttachments["remote_account_1_status_1_attachment_1"]
+	testAttachment.Cached = false
+	err := suite.db.UpdateByPrimaryKey(ctx, testAttachment)
+	suite.NoError(err)
+	err = suite.storage.Delete(testAttachment.File.Path)
+	suite.NoError(err)
+	err = suite.storage.Delete(testAttachment.Thumbnail.Path)
+	suite.NoError(err)
+
+	// now fetch it
+	fileName := path.Base(testAttachment.File.Path)
+	requestingAccount := suite.testAccounts["local_account_1"]
+
+	content, errWithCode := suite.mediaProcessor.GetFile(ctx, requestingAccount, &apimodel.GetContentRequestForm{
+		AccountID: testAttachment.AccountID,
+		MediaType: string(media.TypeAttachment),
+		MediaSize: string(media.SizeOriginal),
+		FileName:  fileName,
+	})
+
+	suite.NoError(errWithCode)
+	suite.NotNil(content)
+	b, err := io.ReadAll(content.Content)
+	suite.NoError(err)
+
+	if closer, ok := content.Content.(io.Closer); ok {
+		suite.NoError(closer.Close())
+	}
+
+	suite.Equal(suite.testRemoteAttachments[testAttachment.RemoteURL].Data, b)
+	suite.Equal(suite.testRemoteAttachments[testAttachment.RemoteURL].ContentType, content.ContentType)
+	suite.EqualValues(len(suite.testRemoteAttachments[testAttachment.RemoteURL].Data), content.ContentLength)
+	time.Sleep(2 * time.Second) // wait a few seconds for the media manager to finish doing stuff
+
+	// the attachment should be updated in the database
+	dbAttachment, err := suite.db.GetAttachmentByID(ctx, testAttachment.ID)
+	suite.NoError(err)
+	suite.True(dbAttachment.Cached)
+
+	// the file should be back in storage at the same path as before
+	refreshedBytes, err := suite.storage.Get(testAttachment.File.Path)
+	suite.NoError(err)
+	suite.Equal(suite.testRemoteAttachments[testAttachment.RemoteURL].Data, refreshedBytes)
+}
+
+func (suite *GetFileTestSuite) TestGetRemoteFileUncachedInterrupted() {
+	ctx := context.Background()
+
+	// uncache the file from local
+	testAttachment := suite.testAttachments["remote_account_1_status_1_attachment_1"]
+	testAttachment.Cached = false
+	err := suite.db.UpdateByPrimaryKey(ctx, testAttachment)
+	suite.NoError(err)
+	err = suite.storage.Delete(testAttachment.File.Path)
+	suite.NoError(err)
+	err = suite.storage.Delete(testAttachment.Thumbnail.Path)
+	suite.NoError(err)
+
+	// now fetch it
+	fileName := path.Base(testAttachment.File.Path)
+	requestingAccount := suite.testAccounts["local_account_1"]
+
+	content, errWithCode := suite.mediaProcessor.GetFile(ctx, requestingAccount, &apimodel.GetContentRequestForm{
+		AccountID: testAttachment.AccountID,
+		MediaType: string(media.TypeAttachment),
+		MediaSize: string(media.SizeOriginal),
+		FileName:  fileName,
+	})
+
+	suite.NoError(errWithCode)
+	suite.NotNil(content)
+
+	// only read the first kilobyte and then stop
+	b := make([]byte, 1024)
+	_, err = content.Content.Read(b)
+	suite.NoError(err)
+
+	// close the reader
+	if closer, ok := content.Content.(io.Closer); ok {
+		suite.NoError(closer.Close())
+	}
+
+	time.Sleep(2 * time.Second) // wait a few seconds for the media manager to finish doing stuff
+
+	// the attachment should still be updated in the database even though the caller hung up
+	dbAttachment, err := suite.db.GetAttachmentByID(ctx, testAttachment.ID)
+	suite.NoError(err)
+	suite.True(dbAttachment.Cached)
+
+	// the file should be back in storage at the same path as before
+	refreshedBytes, err := suite.storage.Get(testAttachment.File.Path)
+	suite.NoError(err)
+	suite.Equal(suite.testRemoteAttachments[testAttachment.RemoteURL].Data, refreshedBytes)
+}
+
+func (suite *GetFileTestSuite) TestGetRemoteFileThumbnailUncached() {
+	ctx := context.Background()
+	testAttachment := suite.testAttachments["remote_account_1_status_1_attachment_1"]
+
+	// fetch the existing thumbnail bytes from storage first
+	thumbnailBytes, err := suite.storage.Get(testAttachment.Thumbnail.Path)
+	suite.NoError(err)
+
+	// uncache the file from local
+	testAttachment.Cached = false
+	err = suite.db.UpdateByPrimaryKey(ctx, testAttachment)
+	suite.NoError(err)
+	err = suite.storage.Delete(testAttachment.File.Path)
+	suite.NoError(err)
+	err = suite.storage.Delete(testAttachment.Thumbnail.Path)
+	suite.NoError(err)
+
+	// now fetch the thumbnail
+	fileName := path.Base(testAttachment.File.Path)
+	requestingAccount := suite.testAccounts["local_account_1"]
+
+	content, errWithCode := suite.mediaProcessor.GetFile(ctx, requestingAccount, &apimodel.GetContentRequestForm{
+		AccountID: testAttachment.AccountID,
+		MediaType: string(media.TypeAttachment),
+		MediaSize: string(media.SizeSmall),
+		FileName:  fileName,
+	})
+
+	suite.NoError(errWithCode)
+	suite.NotNil(content)
+	b, err := io.ReadAll(content.Content)
+	suite.NoError(err)
+
+	if closer, ok := content.Content.(io.Closer); ok {
+		suite.NoError(closer.Close())
+	}
+
+	suite.Equal(thumbnailBytes, b)
+	suite.Equal("image/jpeg", content.ContentType)
+	suite.EqualValues(testAttachment.Thumbnail.FileSize, content.ContentLength)
+}
+
+func TestGetFileTestSuite(t *testing.T) {
+	suite.Run(t, &GetFileTestSuite{})
+}
--- a/internal/processing/media/media.go
+++ b/internal/processing/media/media.go
@@ -27,6 +27,7 @@ import (
 	"github.com/superseriousbusiness/gotosocial/internal/gtserror"
 	"github.com/superseriousbusiness/gotosocial/internal/gtsmodel"
 	"github.com/superseriousbusiness/gotosocial/internal/media"
+	"github.com/superseriousbusiness/gotosocial/internal/transport"
 	"github.com/superseriousbusiness/gotosocial/internal/typeutils"
 )

@@ -36,24 +37,27 @@ type Processor interface {
 	Create(ctx context.Context, account *gtsmodel.Account, form *apimodel.AttachmentRequest) (*apimodel.Attachment, error)
 	// Delete deletes the media attachment with the given ID, including all files pertaining to that attachment.
 	Delete(ctx context.Context, mediaAttachmentID string) gtserror.WithCode
-	GetFile(ctx context.Context, account *gtsmodel.Account, form *apimodel.GetContentRequestForm) (*apimodel.Content, error)
+	// GetFile retrieves a file from storage and streams it back to the caller via an io.reader embedded in *apimodel.Content.
+	GetFile(ctx context.Context, account *gtsmodel.Account, form *apimodel.GetContentRequestForm) (*apimodel.Content, gtserror.WithCode)
 	GetMedia(ctx context.Context, account *gtsmodel.Account, mediaAttachmentID string) (*apimodel.Attachment, gtserror.WithCode)
 	Update(ctx context.Context, account *gtsmodel.Account, mediaAttachmentID string, form *apimodel.AttachmentUpdateRequest) (*apimodel.Attachment, gtserror.WithCode)
 }

 type processor struct {
-	tc           typeutils.TypeConverter
-	mediaManager media.Manager
-	storage      *kv.KVStore
-	db           db.DB
+	tc                  typeutils.TypeConverter
+	mediaManager        media.Manager
+	transportController transport.Controller
+	storage             *kv.KVStore
+	db                  db.DB
 }

 // New returns a new media processor.
-func New(db db.DB, tc typeutils.TypeConverter, mediaManager media.Manager, storage *kv.KVStore) Processor {
+func New(db db.DB, tc typeutils.TypeConverter, mediaManager media.Manager, transportController transport.Controller, storage *kv.KVStore) Processor {
 	return &processor{
-		tc:           tc,
-		mediaManager: mediaManager,
-		storage:      storage,
-		db:           db,
+		tc:                  tc,
+		mediaManager:        mediaManager,
+		transportController: transportController,
+		storage:             storage,
+		db:                  db,
 	}
 }
--- a/internal/processing/media/media_test.go
+++ b/internal/processing/media/media_test.go
@@ -0,0 +1,125 @@
+/*
+   GoToSocial
+   Copyright (C) 2021-2022 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 media_test
+
+import (
+	"bytes"
+	"io"
+	"net/http"
+
+	"codeberg.org/gruf/go-store/kv"
+	"github.com/sirupsen/logrus"
+	"github.com/stretchr/testify/suite"
+	"github.com/superseriousbusiness/gotosocial/internal/db"
+	"github.com/superseriousbusiness/gotosocial/internal/gtsmodel"
+	"github.com/superseriousbusiness/gotosocial/internal/media"
+	mediaprocessing "github.com/superseriousbusiness/gotosocial/internal/processing/media"
+	"github.com/superseriousbusiness/gotosocial/internal/transport"
+	"github.com/superseriousbusiness/gotosocial/internal/typeutils"
+	"github.com/superseriousbusiness/gotosocial/testrig"
+)
+
+type MediaStandardTestSuite struct {
+	// standard suite interfaces
+	suite.Suite
+	db                  db.DB
+	tc                  typeutils.TypeConverter
+	storage             *kv.KVStore
+	mediaManager        media.Manager
+	transportController transport.Controller
+
+	// standard suite models
+	testTokens            map[string]*gtsmodel.Token
+	testClients           map[string]*gtsmodel.Client
+	testApplications      map[string]*gtsmodel.Application
+	testUsers             map[string]*gtsmodel.User
+	testAccounts          map[string]*gtsmodel.Account
+	testAttachments       map[string]*gtsmodel.MediaAttachment
+	testStatuses          map[string]*gtsmodel.Status
+	testRemoteAttachments map[string]testrig.RemoteAttachmentFile
+
+	// module being tested
+	mediaProcessor mediaprocessing.Processor
+}
+
+func (suite *MediaStandardTestSuite) SetupSuite() {
+	suite.testTokens = testrig.NewTestTokens()
+	suite.testClients = testrig.NewTestClients()
+	suite.testApplications = testrig.NewTestApplications()
+	suite.testUsers = testrig.NewTestUsers()
+	suite.testAccounts = testrig.NewTestAccounts()
+	suite.testAttachments = testrig.NewTestAttachments()
+	suite.testStatuses = testrig.NewTestStatuses()
+	suite.testRemoteAttachments = testrig.NewTestFediAttachments("../../../testrig/media")
+}
+
+func (suite *MediaStandardTestSuite) SetupTest() {
+	testrig.InitTestConfig()
+	testrig.InitTestLog()
+
+	suite.db = testrig.NewTestDB()
+	suite.tc = testrig.NewTestTypeConverter(suite.db)
+	suite.storage = testrig.NewTestStorage()
+	suite.mediaManager = testrig.NewTestMediaManager(suite.db, suite.storage)
+	suite.transportController = suite.mockTransportController()
+	suite.mediaProcessor = mediaprocessing.New(suite.db, suite.tc, suite.mediaManager, suite.transportController, suite.storage)
+	testrig.StandardDBSetup(suite.db, nil)
+	testrig.StandardStorageSetup(suite.storage, "../../../testrig/media")
+}
+
+func (suite *MediaStandardTestSuite) TearDownTest() {
+	testrig.StandardDBTeardown(suite.db)
+	testrig.StandardStorageTeardown(suite.storage)
+}
+
+func (suite *MediaStandardTestSuite) mockTransportController() transport.Controller {
+	do := func(req *http.Request) (*http.Response, error) {
+		logrus.Debugf("received request for %s", req.URL)
+
+		responseBytes := []byte{}
+		responseType := ""
+		responseLength := 0
+
+		if attachment, ok := suite.testRemoteAttachments[req.URL.String()]; ok {
+			responseBytes = attachment.Data
+			responseType = attachment.ContentType
+		}
+
+		if len(responseBytes) != 0 {
+			// we found something, so print what we're going to return
+			logrus.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)
+}
--- a/internal/processing/processor.go
+++ b/internal/processing/processor.go
@@ -115,7 +115,7 @@ type Processor interface {
 	BlocksGet(ctx context.Context, authed *oauth.Auth, maxID string, sinceID string, limit int) (*apimodel.BlocksResponse, gtserror.WithCode)

 	// FileGet handles the fetching of a media attachment file via the fileserver.
-	FileGet(ctx context.Context, authed *oauth.Auth, form *apimodel.GetContentRequestForm) (*apimodel.Content, error)
+	FileGet(ctx context.Context, authed *oauth.Auth, form *apimodel.GetContentRequestForm) (*apimodel.Content, gtserror.WithCode)

 	// FollowRequestsGet handles the getting of the authed account's incoming follow requests
 	FollowRequestsGet(ctx context.Context, auth *oauth.Auth) ([]apimodel.Account, gtserror.WithCode)
@@ -270,7 +270,7 @@ func NewProcessor(
 	streamingProcessor := streaming.New(db, oauthServer)
 	accountProcessor := account.New(db, tc, mediaManager, oauthServer, fromClientAPI, federator)
 	adminProcessor := admin.New(db, tc, mediaManager, fromClientAPI)
-	mediaProcessor := mediaProcessor.New(db, tc, mediaManager, storage)
+	mediaProcessor := mediaProcessor.New(db, tc, mediaManager, federator.TransportController(), storage)
 	userProcessor := user.New(db, emailSender)
 	federationProcessor := federationProcessor.New(db, tc, federator, fromFederator)
 	filter := visibility.NewFilter(db)
--- a/internal/processing/streaming/notification_test.go
+++ b/internal/processing/streaming/notification_test.go
@@ -52,7 +52,7 @@ func (suite *NotificationTestSuite) TestStreamNotification() {
 	suite.NoError(err)

 	msg := <-openStream.Messages
-	suite.Equal(`{"id":"01FH57SJCMDWQGEAJ0X08CE3WV","type":"follow","created_at":"2021-10-04T10:52:36+02:00","account":{"id":"01F8MH5ZK5VRH73AKHQM6Y9VNX","username":"foss_satan","acct":"foss_satan@fossbros-anonymous.io","display_name":"big gerald","locked":false,"bot":false,"created_at":"2021-09-26T12:52:36+02:00","note":"i post about like, i dunno, stuff, or whatever!!!!","url":"http://fossbros-anonymous.io/@foss_satan","avatar":"","avatar_static":"","header":"","header_static":"","followers_count":0,"following_count":0,"statuses_count":0,"last_status_at":"","emojis":[],"fields":[]}}`, msg.Payload)
+	suite.Equal(`{"id":"01FH57SJCMDWQGEAJ0X08CE3WV","type":"follow","created_at":"2021-10-04T10:52:36+02:00","account":{"id":"01F8MH5ZK5VRH73AKHQM6Y9VNX","username":"foss_satan","acct":"foss_satan@fossbros-anonymous.io","display_name":"big gerald","locked":false,"bot":false,"created_at":"2021-09-26T12:52:36+02:00","note":"i post about like, i dunno, stuff, or whatever!!!!","url":"http://fossbros-anonymous.io/@foss_satan","avatar":"","avatar_static":"","header":"","header_static":"","followers_count":0,"following_count":0,"statuses_count":1,"last_status_at":"2021-09-20T10:40:37Z","emojis":[],"fields":[]}}`, msg.Payload)
 }

 func TestNotificationTestSuite(t *testing.T) {