Remote instance dereferencing (#70)

Remote instances are now dereferenced when they post to an inbox on a GtS instance.

    Dereferencing will be done first by checking the /api/v1/instance endpoint of an instance.
    If that doesn't work, /.well-known/nodeinfo will be checked.
    If that doesn't work, only a minimal representation of the instance will be stored.

A new field was added to the Instance database model. To create it:

alter table instances add column contact_account_username text;

internal/transport/deliver.go

@@ -0,0 +1,16 @@
+package transport
+
+import (
+	"context"
+	"net/url"
+)
+
+func (t *transport) BatchDeliver(c context.Context, b []byte, recipients []*url.URL) error {
+	return t.sigTransport.BatchDeliver(c, b, recipients)
+}
+
+func (t *transport) Deliver(c context.Context, b []byte, to *url.URL) error {
+	l := t.log.WithField("func", "Deliver")
+	l.Debugf("performing POST to %s", to.String())
+	return t.sigTransport.Deliver(c, b, to)
+}

internal/transport/dereference.go

@@ -0,0 +1,12 @@
+package transport
+
+import (
+	"context"
+	"net/url"
+)
+
+func (t *transport) Dereference(c context.Context, iri *url.URL) ([]byte, error) {
+	l := t.log.WithField("func", "Dereference")
+	l.Debugf("performing GET to %s", iri.String())
+	return t.sigTransport.Dereference(c, iri)
+}

internal/transport/derefinstance.go

@@ -0,0 +1,326 @@
+package transport
+
+import (
+	"context"
+	"encoding/json"
+	"errors"
+	"fmt"
+	"io/ioutil"
+	"net/http"
+	"net/url"
+	"strings"
+
+	apimodel "github.com/superseriousbusiness/gotosocial/internal/api/model"
+	"github.com/superseriousbusiness/gotosocial/internal/gtsmodel"
+	"github.com/superseriousbusiness/gotosocial/internal/id"
+	"github.com/superseriousbusiness/gotosocial/internal/util"
+)
+
+func (t *transport) DereferenceInstance(c context.Context, iri *url.URL) (*gtsmodel.Instance, error) {
+	l := t.log.WithField("func", "DereferenceInstance")
+
+	var i *gtsmodel.Instance
+	var err error
+
+	// First try to dereference using /api/v1/instance.
+	// This will provide the most complete picture of an instance, and avoid unnecessary api calls.
+	//
+	// This will only work with Mastodon-api compatible instances: Mastodon, some Pleroma instances, GoToSocial.
+	l.Debugf("trying to dereference instance %s by /api/v1/instance", iri.Host)
+	i, err = dereferenceByAPIV1Instance(c, t, iri)
+	if err == nil {
+		l.Debugf("successfully dereferenced instance using /api/v1/instance")
+		return i, nil
+	}
+	l.Debugf("couldn't dereference instance using /api/v1/instance: %s", err)
+
+	// If that doesn't work, try to dereference using /.well-known/nodeinfo.
+	// This will involve two API calls and return less info overall, but should be more widely compatible.
+	l.Debugf("trying to dereference instance %s by /.well-known/nodeinfo", iri.Host)
+	i, err = dereferenceByNodeInfo(c, t, iri)
+	if err == nil {
+		l.Debugf("successfully dereferenced instance using /.well-known/nodeinfo")
+		return i, nil
+	}
+	l.Debugf("couldn't dereference instance using /.well-known/nodeinfo: %s", err)
+
+	// we couldn't dereference the instance using any of the known methods, so just return a minimal representation
+	l.Debugf("returning minimal representation of instance %s", iri.Host)
+	id, err := id.NewRandomULID()
+	if err != nil {
+		return nil, fmt.Errorf("error creating new id for instance %s: %s", iri.Host, err)
+	}
+
+	return &gtsmodel.Instance{
+		ID:     id,
+		Domain: iri.Host,
+		URI:    iri.String(),
+	}, nil
+}
+
+func dereferenceByAPIV1Instance(c context.Context, t *transport, iri *url.URL) (*gtsmodel.Instance, error) {
+	l := t.log.WithField("func", "dereferenceByAPIV1Instance")
+
+	cleanIRI := &url.URL{
+		Scheme: iri.Scheme,
+		Host:   iri.Host,
+		Path:   "api/v1/instance",
+	}
+
+	l.Debugf("performing GET to %s", cleanIRI.String())
+	req, err := http.NewRequest("GET", cleanIRI.String(), nil)
+	if err != nil {
+		return nil, err
+	}
+	req = req.WithContext(c)
+	req.Header.Add("Accept", "application/json")
+	req.Header.Add("Date", t.clock.Now().UTC().Format("Mon, 02 Jan 2006 15:04:05")+" GMT")
+	req.Header.Add("User-Agent", fmt.Sprintf("%s %s", t.appAgent, t.gofedAgent))
+	req.Header.Set("Host", cleanIRI.Host)
+	t.getSignerMu.Lock()
+	err = t.getSigner.SignRequest(t.privkey, t.pubKeyID, req, nil)
+	t.getSignerMu.Unlock()
+	if err != nil {
+		return nil, err
+	}
+	resp, err := t.client.Do(req)
+	if err != nil {
+		return nil, err
+	}
+	defer resp.Body.Close()
+	if resp.StatusCode != http.StatusOK {
+		return nil, fmt.Errorf("GET request to %s failed (%d): %s", cleanIRI.String(), resp.StatusCode, resp.Status)
+	}
+	b, err := ioutil.ReadAll(resp.Body)
+	if err != nil {
+		return nil, err
+	}
+
+	if len(b) == 0 {
+		return nil, errors.New("response bytes was len 0")
+	}
+
+	// try to parse the returned bytes directly into an Instance model
+	apiResp := &apimodel.Instance{}
+	if err := json.Unmarshal(b, apiResp); err != nil {
+		return nil, err
+	}
+
+	var contactUsername string
+	if apiResp.ContactAccount != nil {
+		contactUsername = apiResp.ContactAccount.Username
+	}
+
+	ulid, err := id.NewRandomULID()
+	if err != nil {
+		return nil, err
+	}
+
+	i := &gtsmodel.Instance{
+		ID:                     ulid,
+		Domain:                 iri.Host,
+		Title:                  apiResp.Title,
+		URI:                    fmt.Sprintf("%s://%s", iri.Scheme, iri.Host),
+		ShortDescription:       apiResp.ShortDescription,
+		Description:            apiResp.Description,
+		ContactEmail:           apiResp.Email,
+		ContactAccountUsername: contactUsername,
+		Version:                apiResp.Version,
+	}
+
+	return i, nil
+}
+
+func dereferenceByNodeInfo(c context.Context, t *transport, iri *url.URL) (*gtsmodel.Instance, error) {
+	niIRI, err := callNodeInfoWellKnown(c, t, iri)
+	if err != nil {
+		return nil, fmt.Errorf("dereferenceByNodeInfo: error during initial call to well-known nodeinfo: %s", err)
+	}
+
+	ni, err := callNodeInfo(c, t, niIRI)
+	if err != nil {
+		return nil, fmt.Errorf("dereferenceByNodeInfo: error doing second call to nodeinfo uri %s: %s", niIRI.String(), err)
+	}
+
+	// we got a response of some kind! take what we can from it...
+	id, err := id.NewRandomULID()
+	if err != nil {
+		return nil, fmt.Errorf("dereferenceByNodeInfo: error creating new id for instance %s: %s", iri.Host, err)
+	}
+
+	// this is the bare minimum instance we'll return, and we'll add more stuff to it if we can
+	i := &gtsmodel.Instance{
+		ID:     id,
+		Domain: iri.Host,
+		URI:    iri.String(),
+	}
+
+	var title string
+	if i, present := ni.Metadata["nodeName"]; present {
+		// it's present, check it's a string
+		if v, ok := i.(string); ok {
+			// it is a string!
+			title = v
+		}
+	}
+	i.Title = title
+
+	var shortDescription string
+	if i, present := ni.Metadata["nodeDescription"]; present {
+		// it's present, check it's a string
+		if v, ok := i.(string); ok {
+			// it is a string!
+			shortDescription = v
+		}
+	}
+	i.ShortDescription = shortDescription
+
+	var contactEmail string
+	var contactAccountUsername string
+	if i, present := ni.Metadata["maintainer"]; present {
+		// it's present, check it's a map
+		if v, ok := i.(map[string]string); ok {
+			// see if there's an email in the map
+			if email, present := v["email"]; present {
+				if err := util.ValidateEmail(email); err == nil {
+					// valid email address
+					contactEmail = email
+				}
+			}
+			// see if there's a 'name' in the map
+			if name, present := v["name"]; present {
+				// name could be just a username, or could be a mention string eg @whatever@aaaa.com
+				username, _, err := util.ExtractMentionParts(name)
+				if err == nil {
+					// it was a mention string
+					contactAccountUsername = username
+				} else {
+					// not a mention string
+					contactAccountUsername = name
+				}
+			}
+		}
+	}
+	i.ContactEmail = contactEmail
+	i.ContactAccountUsername = contactAccountUsername
+
+	var software string
+	if ni.Software.Name != "" {
+		software = ni.Software.Name
+	}
+	if ni.Software.Version != "" {
+		software = software + " " + ni.Software.Version
+	}
+	i.Version = software
+
+	return i, nil
+}
+
+func callNodeInfoWellKnown(c context.Context, t *transport, iri *url.URL) (*url.URL, error) {
+	l := t.log.WithField("func", "callNodeInfoWellKnown")
+
+	cleanIRI := &url.URL{
+		Scheme: iri.Scheme,
+		Host:   iri.Host,
+		Path:   ".well-known/nodeinfo",
+	}
+
+	l.Debugf("performing GET to %s", cleanIRI.String())
+	req, err := http.NewRequest("GET", cleanIRI.String(), nil)
+	if err != nil {
+		return nil, err
+	}
+	req = req.WithContext(c)
+	req.Header.Add("Accept", "application/json")
+	req.Header.Add("Date", t.clock.Now().UTC().Format("Mon, 02 Jan 2006 15:04:05")+" GMT")
+	req.Header.Add("User-Agent", fmt.Sprintf("%s %s", t.appAgent, t.gofedAgent))
+	req.Header.Set("Host", cleanIRI.Host)
+	t.getSignerMu.Lock()
+	err = t.getSigner.SignRequest(t.privkey, t.pubKeyID, req, nil)
+	t.getSignerMu.Unlock()
+	if err != nil {
+		return nil, err
+	}
+	resp, err := t.client.Do(req)
+	if err != nil {
+		return nil, err
+	}
+	defer resp.Body.Close()
+	if resp.StatusCode != http.StatusOK {
+		return nil, fmt.Errorf("callNodeInfoWellKnown: GET request to %s failed (%d): %s", cleanIRI.String(), resp.StatusCode, resp.Status)
+	}
+	b, err := ioutil.ReadAll(resp.Body)
+	if err != nil {
+		return nil, err
+	}
+
+	if len(b) == 0 {
+		return nil, errors.New("callNodeInfoWellKnown: response bytes was len 0")
+	}
+
+	wellKnownResp := &apimodel.WellKnownResponse{}
+	if err := json.Unmarshal(b, wellKnownResp); err != nil {
+		return nil, fmt.Errorf("callNodeInfoWellKnown: could not unmarshal server response as WellKnownResponse: %s", err)
+	}
+
+	// look through the links for the first one that matches the nodeinfo schema, this is what we need
+	var nodeinfoHref *url.URL
+	for _, l := range wellKnownResp.Links {
+		if l.Href == "" || !strings.HasPrefix(l.Rel, "http://nodeinfo.diaspora.software/ns/schema/2") {
+			continue
+		}
+		nodeinfoHref, err = url.Parse(l.Href)
+		if err != nil {
+			return nil, fmt.Errorf("callNodeInfoWellKnown: couldn't parse url %s: %s", l.Href, err)
+		}
+	}
+	if nodeinfoHref == nil {
+		return nil, errors.New("callNodeInfoWellKnown: could not find nodeinfo rel in well known response")
+	}
+
+	return nodeinfoHref, nil
+}
+
+func callNodeInfo(c context.Context, t *transport, iri *url.URL) (*apimodel.Nodeinfo, error) {
+	l := t.log.WithField("func", "callNodeInfo")
+
+	l.Debugf("performing GET to %s", iri.String())
+	req, err := http.NewRequest("GET", iri.String(), nil)
+	if err != nil {
+		return nil, err
+	}
+	req = req.WithContext(c)
+	req.Header.Add("Accept", "application/json")
+	req.Header.Add("Date", t.clock.Now().UTC().Format("Mon, 02 Jan 2006 15:04:05")+" GMT")
+	req.Header.Add("User-Agent", fmt.Sprintf("%s %s", t.appAgent, t.gofedAgent))
+	req.Header.Set("Host", iri.Host)
+	t.getSignerMu.Lock()
+	err = t.getSigner.SignRequest(t.privkey, t.pubKeyID, req, nil)
+	t.getSignerMu.Unlock()
+	if err != nil {
+		return nil, err
+	}
+	resp, err := t.client.Do(req)
+	if err != nil {
+		return nil, err
+	}
+	defer resp.Body.Close()
+	if resp.StatusCode != http.StatusOK {
+		return nil, fmt.Errorf("callNodeInfo: GET request to %s failed (%d): %s", iri.String(), resp.StatusCode, resp.Status)
+	}
+	b, err := ioutil.ReadAll(resp.Body)
+	if err != nil {
+		return nil, err
+	}
+
+	if len(b) == 0 {
+		return nil, errors.New("callNodeInfo: response bytes was len 0")
+	}
+
+	niResp := &apimodel.Nodeinfo{}
+	if err := json.Unmarshal(b, niResp); err != nil {
+		return nil, fmt.Errorf("callNodeInfo: could not unmarshal server response as Nodeinfo: %s", err)
+	}
+
+	return niResp, nil
+}

internal/transport/derefmedia.go

@@ -0,0 +1,42 @@
+package transport
+
+import (
+	"context"
+	"fmt"
+	"io/ioutil"
+	"net/http"
+	"net/url"
+)
+
+func (t *transport) DereferenceMedia(c context.Context, iri *url.URL, expectedContentType string) ([]byte, error) {
+	l := t.log.WithField("func", "DereferenceMedia")
+	l.Debugf("performing GET to %s", iri.String())
+	req, err := http.NewRequest("GET", iri.String(), nil)
+	if err != nil {
+		return nil, err
+	}
+	req = req.WithContext(c)
+	if expectedContentType == "" {
+		req.Header.Add("Accept", "*/*")
+	} else {
+		req.Header.Add("Accept", expectedContentType)
+	}
+	req.Header.Add("Date", t.clock.Now().UTC().Format("Mon, 02 Jan 2006 15:04:05")+" GMT")
+	req.Header.Add("User-Agent", fmt.Sprintf("%s %s", t.appAgent, t.gofedAgent))
+	req.Header.Set("Host", iri.Host)
+	t.getSignerMu.Lock()
+	err = t.getSigner.SignRequest(t.privkey, t.pubKeyID, req, nil)
+	t.getSignerMu.Unlock()
+	if err != nil {
+		return nil, err
+	}
+	resp, err := t.client.Do(req)
+	if err != nil {
+		return nil, err
+	}
+	defer resp.Body.Close()
+	if resp.StatusCode != http.StatusOK {
+		return nil, fmt.Errorf("GET request to %s failed (%d): %s", iri.String(), resp.StatusCode, resp.Status)
+	}
+	return ioutil.ReadAll(resp.Body)
+}

internal/transport/finger.go

@@ -0,0 +1,48 @@
+package transport
+
+import (
+	"context"
+	"fmt"
+	"io/ioutil"
+	"net/http"
+	"net/url"
+)
+
+func (t *transport) Finger(c context.Context, targetUsername string, targetDomain string) ([]byte, error) {
+	l := t.log.WithField("func", "Finger")
+	urlString := fmt.Sprintf("https://%s/.well-known/webfinger?resource=acct:%s@%s", targetDomain, targetUsername, targetDomain)
+	l.Debugf("performing GET to %s", urlString)
+
+	iri, err := url.Parse(urlString)
+	if err != nil {
+		return nil, fmt.Errorf("Finger: error parsing url %s: %s", urlString, err)
+	}
+
+	l.Debugf("performing GET to %s", iri.String())
+
+	req, err := http.NewRequest("GET", iri.String(), nil)
+	if err != nil {
+		return nil, err
+	}
+	req = req.WithContext(c)
+	req.Header.Add("Accept", "application/json")
+	req.Header.Add("Accept", "application/jrd+json")
+	req.Header.Add("Date", t.clock.Now().UTC().Format("Mon, 02 Jan 2006 15:04:05")+" GMT")
+	req.Header.Add("User-Agent", fmt.Sprintf("%s %s", t.appAgent, t.gofedAgent))
+	req.Header.Set("Host", iri.Host)
+	t.getSignerMu.Lock()
+	err = t.getSigner.SignRequest(t.privkey, t.pubKeyID, req, nil)
+	t.getSignerMu.Unlock()
+	if err != nil {
+		return nil, err
+	}
+	resp, err := t.client.Do(req)
+	if err != nil {
+		return nil, err
+	}
+	defer resp.Body.Close()
+	if resp.StatusCode != http.StatusOK {
+		return nil, fmt.Errorf("GET request to %s failed (%d): %s", iri.String(), resp.StatusCode, resp.Status)
+	}
+	return ioutil.ReadAll(resp.Body)
+}

internal/transport/transport.go

@@ -3,22 +3,23 @@ package transport
 import (
 	"context"
 	"crypto"
-	"fmt"
-	"io/ioutil"
-	"net/http"
 	"net/url"
 	"sync"
 
 	"github.com/go-fed/activity/pub"
 	"github.com/go-fed/httpsig"
 	"github.com/sirupsen/logrus"
+	"github.com/superseriousbusiness/gotosocial/internal/gtsmodel"
 )
 
 // Transport wraps the pub.Transport interface with some additional
 // functionality for fetching remote media.
 type Transport interface {
 	pub.Transport
+	// DereferenceMedia fetches the bytes of the given media attachment IRI, with the expectedContentType.
 	DereferenceMedia(c context.Context, iri *url.URL, expectedContentType string) ([]byte, error)
+	// DereferenceInstance dereferences remote instance information, first by checking /api/v1/instance, and then by checking /.well-known/nodeinfo.
+	DereferenceInstance(c context.Context, iri *url.URL) (*gtsmodel.Instance, error)
 	// Finger performs a webfinger request with the given username and domain, and returns the bytes from the response body.
 	Finger(c context.Context, targetUsername string, targetDomains string) ([]byte, error)
 }
@@ -36,91 +37,3 @@ type transport struct {
 	getSignerMu  *sync.Mutex
 	log          *logrus.Logger
 }
-
-func (t *transport) BatchDeliver(c context.Context, b []byte, recipients []*url.URL) error {
-	return t.sigTransport.BatchDeliver(c, b, recipients)
-}
-
-func (t *transport) Deliver(c context.Context, b []byte, to *url.URL) error {
-	l := t.log.WithField("func", "Deliver")
-	l.Debugf("performing POST to %s", to.String())
-	return t.sigTransport.Deliver(c, b, to)
-}
-
-func (t *transport) Dereference(c context.Context, iri *url.URL) ([]byte, error) {
-	l := t.log.WithField("func", "Dereference")
-	l.Debugf("performing GET to %s", iri.String())
-	return t.sigTransport.Dereference(c, iri)
-}
-
-func (t *transport) DereferenceMedia(c context.Context, iri *url.URL, expectedContentType string) ([]byte, error) {
-	l := t.log.WithField("func", "DereferenceMedia")
-	l.Debugf("performing GET to %s", iri.String())
-	req, err := http.NewRequest("GET", iri.String(), nil)
-	if err != nil {
-		return nil, err
-	}
-	req = req.WithContext(c)
-	if expectedContentType == "" {
-		req.Header.Add("Accept", "*/*")
-	} else {
-		req.Header.Add("Accept", expectedContentType)
-	}
-	req.Header.Add("Date", t.clock.Now().UTC().Format("Mon, 02 Jan 2006 15:04:05")+" GMT")
-	req.Header.Add("User-Agent", fmt.Sprintf("%s %s", t.appAgent, t.gofedAgent))
-	req.Header.Set("Host", iri.Host)
-	t.getSignerMu.Lock()
-	err = t.getSigner.SignRequest(t.privkey, t.pubKeyID, req, nil)
-	t.getSignerMu.Unlock()
-	if err != nil {
-		return nil, err
-	}
-	resp, err := t.client.Do(req)
-	if err != nil {
-		return nil, err
-	}
-	defer resp.Body.Close()
-	if resp.StatusCode != http.StatusOK {
-		return nil, fmt.Errorf("GET request to %s failed (%d): %s", iri.String(), resp.StatusCode, resp.Status)
-	}
-	return ioutil.ReadAll(resp.Body)
-}
-
-func (t *transport) Finger(c context.Context, targetUsername string, targetDomain string) ([]byte, error) {
-	l := t.log.WithField("func", "Finger")
-	urlString := fmt.Sprintf("https://%s/.well-known/webfinger?resource=acct:%s@%s", targetDomain, targetUsername, targetDomain)
-	l.Debugf("performing GET to %s", urlString)
-
-	iri, err := url.Parse(urlString)
-	if err != nil {
-		return nil, fmt.Errorf("Finger: error parsing url %s: %s", urlString, err)
-	}
-
-	l.Debugf("performing GET to %s", iri.String())
-
-	req, err := http.NewRequest("GET", iri.String(), nil)
-	if err != nil {
-		return nil, err
-	}
-	req = req.WithContext(c)
-	req.Header.Add("Accept", "application/json")
-	req.Header.Add("Accept", "application/jrd+json")
-	req.Header.Add("Date", t.clock.Now().UTC().Format("Mon, 02 Jan 2006 15:04:05")+" GMT")
-	req.Header.Add("User-Agent", fmt.Sprintf("%s %s", t.appAgent, t.gofedAgent))
-	req.Header.Set("Host", iri.Host)
-	t.getSignerMu.Lock()
-	err = t.getSigner.SignRequest(t.privkey, t.pubKeyID, req, nil)
-	t.getSignerMu.Unlock()
-	if err != nil {
-		return nil, err
-	}
-	resp, err := t.client.Do(req)
-	if err != nil {
-		return nil, err
-	}
-	defer resp.Body.Close()
-	if resp.StatusCode != http.StatusOK {
-		return nil, fmt.Errorf("GET request to %s failed (%d): %s", iri.String(), resp.StatusCode, resp.Status)
-	}
-	return ioutil.ReadAll(resp.Body)
-}