Files
archie/internal/channels/telegram/media.go
T
sam e0b6aa6105 feat(workflow): post line-anchored review comments with suggested fixes (#891)
* feat(workflow): post line-anchored review comments with suggested fixes

Bead archie-core-q9au (docs/prds/inline-review.md). A review's findings reached
the operator only as a flat list in the PR body, so the author had to map every
finding back to its line by hand and apply mechanical fixes by hand. Now the
line-anchored findings are also posted as inline review comments, carrying a
one-click suggestion when the reviewer could state the fix.

Domain (internal/domain/workflow):
- ReviewFinding gains Suggestion; Validate rejects one on anything but a
  confirmed finding, so the agent's record_finding call is refused with
  something it can act on. Validate splits into validateAnchor /
  validateClassification / validateDisposition, which also brings it back under
  the cyclop limit the gate enforces.
- StagePostReviewComments runs after StageOpenPR in the implement workflow. It
  posts only line-anchored findings (Line > 0); a whole-file finding has no line
  to attach to and stays in the body list, which remains the complete record.
  It is best-effort: the PR is open and the body already lists every finding, so
  a failed comment is logged, never a park -- the same reasoning as OpenPR's
  existing best-effort LinkBranch call.
- The suggestion fence is gated on Verdict == confirmed where the fence is
  written, because it is a one-click apply: a wrong suggestion breaks the code
  and the author has to notice.

Forge (internal/forge): ReviewCommentWriter + GitHub/Gitea implementations,
type-asserted like PullRequestReader so a forge that cannot is skipped rather
than faked (NoopForge is pinned as not satisfying it). GitHub loops one
CreateComment per comment and joins the failures, so one refused anchor cannot
drop the rest; Gitea carries them on a single COMMENT-state review.

Two deliberate departures from the PRD's sketch, both recorded in it:
- The interface is batch (CreateReviewComments), not one call per comment:
  Gitea's only inline shape is one submitted review, and the set then crosses
  NATS once instead of once per finding.
- The head SHA is resolved inside each forge implementation rather than by the
  caller. The stage runs in archie-agent, which holds no forge credentials and
  has no PullRequestReader; the daemon-side forgerpc handler is the only place
  that does.

Transport (internal/forgerpc): workflow.Forger gains the method, so the agent
reaches the daemon over the new create_review_comments subject; the server
type-asserts the writer and answers an incapable forge with an error the stage
logs. This is the one proxied method whose absence on an image skew does not
park a task.

The reviewer prompt and record_finding schema ask for a suggestion on a
confirmed, mechanically fixable defect only.

Regenerating the wfextract symbol table -- the documented way a new exported
workflow symbol reaches interpreted stage plugins -- also picked up symbols that
were missing from the committed table (the Status* and Source* constants,
ReviewCheck, Store, Task, and the _Store wrapper). Nil-guards were re-applied to
every wrapper afterwards, as that file's header requires.

Evidence: task check exit 0 (incl. 229 UI tests). Three mutations of the new
stage logic each fail a test that names the behaviour -- dropping the
confirmed-only gate, dropping the Line > 0 filter, and letting a posting failure
escape -- so the tests discriminate rather than passing vacuously.

* fix(build): drop the Dockerfile pin assertion

The guard fetched golangci-lint's go.mod at image build time to fail the build
when GOLANGCI_LINT_VERSION and GOFUMPT_VERSION disagreed. It only ever tripped
on self-inflicted pin drift, but it added a network fetch and a new failure
mode to the release build, which is not where that belongs.

The pins stay correct (golangci-lint 2.13.2, whose vendored gofumpt v0.11.0
equals GOFUMPT_VERSION) and the comment above them still states the invariant,
so the reasoning survives without the build-time machinery. The pin change
itself reaches the agent image through the normal release path: CI rebuilds
and pushes archie-agent on an archie/vX.Y.Z release, and a gateway-only
release deliberately does not.

* feat(workflow): add the remediate workflow for PR review remediation

Implements step 5 of docs/prds/pr-review-remediation.md: a remediate
workflow that resumes the task's existing worktree onto its PR branch
(worktree.Manager.Resume, not Prepare's refresh), runs a builder agent
against the review's actionable comments, commits and pushes to the
same branch, and replies via the forge's ReplyToReview. Task.RetryCount
doubles as the remediation round counter, capped by the repo's
max_retries; the round-cap stage parks the task and posts one comment
when it's hit.

Task gains a ReviewPayload field (JSON-encoded review unit) with full
store/proto/staterpc plumbing. workflow.Forger gains Comment and
ReplyToReview, proxied to archie-agent via a new forgerpc RPC. Resume
and its Dir lookup are added as an unexported optional Trees capability
rather than widening the exported interface, which is already at
golangci-lint's interfacebloat cap and is projected into wfextract's
Yaegi symbol table for repository-authored stage plugins.

The daemon-side reaction consumer that dispatches into this workflow
(subscribing ARCHIE_REACTIONS, grouping comments by review, enforcing
per-task serialization) is tracked separately in archie-core-8li9.4.

* feat(messaging): extract standalone messaging service binary and process composition

* feat(messaging): adapt channel frontends to consume Gateway ChatContract

* feat(messaging): remove in-process channels and Router from daemon bootstrap

Telegram, email and webhook now run only in the Messaging Service, which
reaches this process through the Gateway's ChatContract. The daemon builds
no channel and no gateway.Router, so the wiring that existed only to serve
them goes with it: setupGateways, setupEmailGateway, setupWebhookGateway,
setupTelegramGateway, buildTelegramRouter, makeTelegramReload and the
release-announcement path helper.

The seams those paths owned move to where they can still be honoured:

- Channel ValidateConfig runs in archiemessaging.compose, which now returns
  an error, so a configured-but-invalid front-end fails the service rather
  than being dropped silently.
- The inbound rate limiter moves to the Gateway process, which owns the only
  Router and therefore the only place a per-sender budget applies to every
  channel's turns.
- The daemon's gateway readiness probe becomes a ContractProbe against the
  Gateway. It cannot observe a channel lifecycle it no longer runs, and a
  channel-count probe would have reported every daemon degraded.
- telegramSetup is now chatSetup (in chat_setup.go), trimmed to the fields a
  chat turn runner actually reads.

Telegram operator features that were wired only in the daemon (/update,
release announcements, version reporting, SIGHUP reload) are not yet wired
in the Messaging Service: archie-core-8cda.6.6. Channel lifecycle reporting:
archie-core-8cda.6.8. The Gateway process builds no worktree manager, so
review_pr is unadvertised on every channel: archie-core-8cda.6.7.

* feat(messaging): re-wire the Telegram operator surface in the Messaging Service

setupTelegramGateway satisfied eight telegram.Gateway seams from the daemon's
own process. Each is resolved by where the fact it needs is owned, not by
where it used to be built; the decision table is in
docs/architecture/migration-decisions.md.

Wired here: /update (built from this service's own [chat.telegram] check and
install commands), the update-report relay (same sha256(bot_user)[:8] path the
daemon wrote, so a pending report is found rather than written afresh),
show_tool_calls, and /restart's config reload, which re-resolves the token and
allowlist from this service's own file and refuses a reload that resolves to
an empty token.

/version is sourced from the Gateway: ChatSnapshot gains a Version field
carrying the Gateway's rendered build block, read per call and bounded by the
dependency timeout. cmd/archie-messaging deliberately takes no version
ldflags, so it cannot report a partial upgrade as a matched one.

RunningVersions and ReleaseAnnouncements stay nil. Both need a component's
real running version, and the only versions this process can obtain are a
different binary's stamps -- reporting them would manufacture exactly the
false success the update check exists to catch, and the announcer without
parseable versions is a no-op that reads as configured. archie-core-8cda.6.9
carries the contract that unblocks them. Dangerous stays nil: no composition
has ever supplied a command authority.

Enrich is dropped from the update service: componentInstallTypeEnricher read
[nats], which this service must not decode, so the check command's own
reported install type stands.

* fix(workflow): guard inline review comments against head drift

Two findings from the review of q9au.

A comment's line numbers describe exactly one revision of the branch, but the
forge resolved the pull request's head at posting time -- after the review had
already read the worktree. A push in that window (a collaborator, a retry) meant
the findings were anchored to whatever now occupied those lines, mislabelling
instead of failing. The revision the review measured now travels with the
comment set: OpenPR records it from the same worktree read it already performs
(TaskContext.ReviewedHeadSHA), StagePostReviewComments threads it through
Forger to forgerpc, and the forge refuses the whole set once the head it reads
has moved off it -- which the best-effort stage logs, the PR body remaining the
record. An unmeasured revision posts unverified rather than dropping every
finding, so a worktree that cannot report one degrades instead of regressing.

Suggestion was documented single-line only, because GitHub applies a suggestion
fence up to its first line and a multi-line replacement would silently corrupt
the file. Only the verdict was validated, so a multi-line suggestion reached the
renderer and posted as that truncated fence. Validate now rejects an embedded
newline, which makes record_finding refuse it with feedback the model can act on
instead of the reviewer discovering the corruption later.
2026-09-19 23:37:23 +10:00

249 lines
8.9 KiB
Go

package telegram
import (
"context"
"errors"
"fmt"
"os"
"path/filepath"
"github.com/go-telegram/bot"
"github.com/go-telegram/bot/models"
"github.com/samcharles93/archie-core/internal/domain/messaging"
)
// telegramMediaSender delivers a MessageEvent's attachment through the
// media-specific Bot API methods (sendVideo, sendPhoto, ...) rather than
// sendMessage, which cannot carry a file at all.
//
// It captures the *bot.Bot at construction rather than reading the
// Gateway's shared field, matching telegramApprover: a sender belongs to
// the launch that built it, so a /restart abandons in-flight sends with
// the outgoing bot instance instead of silently redirecting them through
// the new one. That is the same lifetime rule the per-launch turns
// registry documents.
//
// An attachment is delivered one of two ways, chosen by what it carries:
// a URL is handed to Telegram to FETCH, while a Path is UPLOADED from this
// host. Delivery used to be URL-only, which meant a locally produced file
// was passed off as a URL Telegram could not fetch and the send did
// nothing useful while reporting success.
//
// A MediaAttachment.FileID is still not accepted: it identifies a file on
// the platform it was uploaded from and is meaningless as an outbound
// handle here.
type telegramMediaSender struct {
bot *bot.Bot
chatID int64
threadID int
}
var (
_ messaging.MediaSender = (*telegramMediaSender)(nil)
_ messaging.CapabilityReporter = (*telegramMediaSender)(nil)
)
// NewMediaSender returns a MediaSender bound to one chat and to the bot
// instance of the launch that created it. Construct it per launch, from
// the same b that handled the update, never from a stored field.
func (g *Gateway) NewMediaSender(b *bot.Bot, chatID int64, threadID int) messaging.MediaSender {
return &telegramMediaSender{bot: b, chatID: chatID, threadID: threadID}
}
// Capabilities reports Media support. Telegram implements every media
// kind MediaAttachment can describe, so this is unconditionally true.
func (s *telegramMediaSender) Capabilities() messaging.AdapterCapabilities {
return messaging.AdapterCapabilities{Media: true}
}
// errNoMedia and errNoSource are the two ways an event can be
// undeliverable before any request is made.
var (
errNoMedia = errors.New("message event carries no media attachment")
// errNoSource replaces the former errNoURL: an attachment is now
// deliverable with either a URL or a local Path, so having neither --
// not having no URL -- is what makes it undeliverable.
errNoSource = errors.New("media attachment has neither a URL nor a local path")
)
// Bot API upload ceilings, in bytes. Photos are capped far lower than
// everything else, and exceeding either is a 400 from Telegram with a
// message the operator never sees; checking here turns that into a
// reported failure with the actual size in it.
const (
maxPhotoUploadBytes int64 = 10 * 1024 * 1024
maxFileUploadBytes int64 = 50 * 1024 * 1024
)
// SendMedia delivers the event's first attachment, captioned with the
// event text. Only the first is sent: a caption belongs to one file, and
// batching several into a media group is a distinct API with its own
// caption semantics.
func (s *telegramMediaSender) SendMedia(ctx context.Context, event messaging.MessageEvent) (messaging.SendResult, error) {
if len(event.Media) == 0 {
return invalidMedia(errNoMedia)
}
att := event.Media[0]
if att.URL == "" && att.Path == "" {
return invalidMedia(fmt.Errorf("%w (type %q)", errNoSource, att.Type))
}
file, closeFile, err := s.source(att)
if err != nil {
return invalidMedia(err)
}
defer closeFile()
msg, err := s.dispatch(ctx, att, file, event.Text)
if err != nil {
if errors.Is(err, errUnsupportedMediaType) {
return invalidMedia(err)
}
res := classifySendError(err)
res.Error = fmt.Errorf("send %s: %w", att.Type, err)
return res, res.Error
}
id := ""
if msg != nil {
id = fmt.Sprintf("%d", msg.ID)
}
return messaging.SendResult{Success: true, MessageID: id}, nil
}
var (
errUnsupportedMediaType = errors.New("unsupported media type")
errNotRegularFile = errors.New("not a regular file")
errTooLarge = errors.New("file exceeds the Telegram upload limit")
)
// source builds the Bot API file value for att, and a function to release
// whatever it holds open.
//
// A URL becomes an InputFileString, which tells Telegram to fetch it. A
// local path becomes an InputFileUpload carrying an open *os.File, which
// the library streams as multipart form data -- the only way bytes that
// exist solely on this host can reach a chat.
//
// The size check happens here rather than being left to Telegram: over the
// limit the API answers a bad request whose text nobody reads, and the
// operator sees an attachment that never arrived. The returned error is
// classified invalid_message by the caller, so it is not retried -- a
// file does not get smaller on a second attempt.
func (s *telegramMediaSender) source(att messaging.MediaAttachment) (models.InputFile, func(), error) {
noop := func() {}
if att.Path == "" {
return &models.InputFileString{Data: att.URL}, noop, nil
}
info, err := os.Stat(att.Path)
if err != nil {
return nil, noop, fmt.Errorf("attach %s: %w", att.Path, err)
}
if !info.Mode().IsRegular() {
return nil, noop, fmt.Errorf("attach %s: %w", att.Path, errNotRegularFile)
}
if limit := uploadLimit(att.Type); info.Size() > limit {
return nil, noop, fmt.Errorf("attach %s: %w: %d bytes exceeds %d",
att.Path, errTooLarge, info.Size(), limit)
}
f, err := os.Open(att.Path)
if err != nil {
return nil, noop, fmt.Errorf("attach %s: %w", att.Path, err)
}
name := att.FileName
if name == "" {
name = filepath.Base(att.Path)
}
return &models.InputFileUpload{Filename: name, Data: f}, func() { _ = f.Close() }, nil
}
// uploadLimit reports the Bot API ceiling for a media kind. Photos have
// their own, much lower one; everything else shares the general file
// limit.
func uploadLimit(mediaType string) int64 {
if mediaType == "image" {
return maxPhotoUploadBytes
}
return maxFileUploadBytes
}
// dispatch routes an attachment to the Bot API method for its kind. The
// media type, not the MessageEvent type, decides: the event type describes
// the message while the attachment describes the file, and Telegram
// rejects a video sent through sendPhoto.
func (s *telegramMediaSender) dispatch(ctx context.Context, att messaging.MediaAttachment, file models.InputFile, caption string) (*models.Message, error) {
switch att.Type {
case "video":
p := &bot.SendVideoParams{ChatID: s.chatID, Video: file, Caption: caption}
p.MessageThreadID = s.threadID
if att.Width != nil {
p.Width = *att.Width
}
if att.Height != nil {
p.Height = *att.Height
}
if att.Duration != nil {
p.Duration = *att.Duration
}
return s.bot.SendVideo(ctx, p)
case "image":
p := &bot.SendPhotoParams{ChatID: s.chatID, Photo: file, Caption: caption}
p.MessageThreadID = s.threadID
return s.bot.SendPhoto(ctx, p)
case "audio":
p := &bot.SendAudioParams{ChatID: s.chatID, Audio: file, Caption: caption}
p.MessageThreadID = s.threadID
if att.Duration != nil {
p.Duration = *att.Duration
}
return s.bot.SendAudio(ctx, p)
case "document":
p := &bot.SendDocumentParams{ChatID: s.chatID, Document: file, Caption: caption}
p.MessageThreadID = s.threadID
return s.bot.SendDocument(ctx, p)
default:
return nil, fmt.Errorf("%w %q", errUnsupportedMediaType, att.Type)
}
}
// invalidMedia reports a failure that was detected before any request was
// made. Never retryable: resending an event that carries nothing, or names
// a type Telegram has no method for, fails identically every time.
func invalidMedia(err error) (messaging.SendResult, error) {
return messaging.SendResult{
Success: false,
Retryable: false,
Error: err,
ErrorCode: "invalid_message",
}, err
}
// classifySendError maps a Bot API failure onto SendResult's error
// vocabulary so a caller can decide on retry without inspecting
// Telegram-specific errors.
//
// The library reports API failures as sentinel errors (bot.ErrorBadRequest
// and friends), not as typed values, so these are errors.Is checks. A
// failure matching no sentinel is a transport error or a 5xx, both of
// which may yet succeed -- hence retryable, rather than defaulting to
// permanent and silently dropping a deliverable asset.
func classifySendError(err error) messaging.SendResult {
res := messaging.SendResult{Success: false, Error: err}
switch {
case bot.IsTooManyRequestsError(err) || errors.Is(err, bot.ErrorTooManyRequests):
res.Retryable, res.ErrorCode = true, "rate_limited"
case errors.Is(err, bot.ErrorUnauthorized), errors.Is(err, bot.ErrorForbidden):
res.Retryable, res.ErrorCode = false, "auth"
case errors.Is(err, bot.ErrorBadRequest), errors.Is(err, bot.ErrorNotFound):
res.Retryable, res.ErrorCode = false, "invalid_message"
default:
res.Retryable, res.ErrorCode = true, "network"
}
return res
}