From 50cdcc0376bc11e4a95e63a86cee3ee529072bdf Mon Sep 17 00:00:00 2001 From: lymkwi Date: Wed, 1 May 2024 01:07:11 +0200 Subject: [PATCH] Implement attachment captioning Starting with Client-Server API v1.10 [0], the `body` field in messages of type `m.image`, `m.audio`, `m.video` and `m.file` can be used as the caption of the attachment. This is fact the way that Nheko rends captions on images, for example. This commit introduces a field in the `UploadHandle`s awaiting upload on the timeline's `InputBar` which holds a caption taken from the input text area. The decision is as follows: - If text bar is empty or full of blanks, send all media with no caption - If the text is an incomplete command, fail - If there are no pending uploads, proceed as done previously (if there is no command recognized send the text, or try and execute the command and if it fails send the text) - If there are pending uploads, only accept uploads if nothing resembling a command name is in the text area. That text becomes the caption for all pending media. Otherwise, try and execute the command, and, if it fails, send it as text. While this workflow for captioning so far is a bit jank, it is the least effort implementation. Links: [0]: https://spec.matrix.org/v1.10/client-server-api/#mimage Signed-off-by: lymkwi --- src/timeline/InputBar.cpp | 52 ++++++++++++++++++++++++++++++--------- src/timeline/InputBar.h | 7 ++++++ 2 files changed, 48 insertions(+), 11 deletions(-) diff --git a/src/timeline/InputBar.cpp b/src/timeline/InputBar.cpp index 1a086c2b..8cae463e 100644 --- a/src/timeline/InputBar.cpp +++ b/src/timeline/InputBar.cpp @@ -394,6 +394,9 @@ InputBar::send() { QInputMethod *im = QGuiApplication::inputMethod(); im->commit(); + // If the input from the UI is only blanks or no text, this trigger should + // be used to confirm media upload. If that is not the case however, but + // but there are pending uploads, we fall into one of the cases seen later. if (text().trimmed().isEmpty()) { acceptUploads(); return; @@ -407,8 +410,18 @@ InputBar::send() updateTextContentProperties(text()); if (containsIncompleteCommand_) return; - if (commandName.isEmpty() || !command(commandName, args)) - message(text()); + if (unconfirmedUploads.empty()) { + if (commandName.isEmpty() || !command(commandName, args)) { + message(text()); + } + } else { + if (commandName.isEmpty()) { + // This is a set of uploads with text + acceptUploadsWithCaption(text()); + } else if (!command(commandName, args)) { + message(text()); + } + } if (!wasEdit) { history_.push_front(QLatin1String("")); @@ -708,6 +721,7 @@ void InputBar::image(const QString &filename, const std::optional &file, const QString &url, + const std::optional &caption, const QString &mime, uint64_t dsize, const QSize &dimensions, @@ -721,9 +735,10 @@ InputBar::image(const QString &filename, image.info.mimetype = mime.toStdString(); image.info.size = dsize; image.info.blurhash = blurhash.toStdString(); - image.body = filename.toStdString(); - image.info.h = dimensions.height(); - image.info.w = dimensions.width(); + // Depending on the input bar's situation, retrieve the text + image.body = caption.has_value() ? caption.value().toStdString() : filename.toStdString(); + image.info.h = dimensions.height(); + image.info.w = dimensions.width(); if (file) image.file = file; @@ -752,13 +767,14 @@ void InputBar::file(const QString &filename, const std::optional &encryptedFile, const QString &url, + const std::optional &caption, const QString &mime, uint64_t dsize) { mtx::events::msg::File file; file.info.mimetype = mime.toStdString(); file.info.size = dsize; - file.body = filename.toStdString(); + file.body = caption.has_value() ? caption.value().toStdString() : filename.toStdString(); if (encryptedFile) file.file = encryptedFile; @@ -775,6 +791,7 @@ void InputBar::audio(const QString &filename, const std::optional &file, const QString &url, + const std::optional &caption, const QString &mime, uint64_t dsize, uint64_t duration) @@ -782,8 +799,8 @@ InputBar::audio(const QString &filename, mtx::events::msg::Audio audio; audio.info.mimetype = mime.toStdString(); audio.info.size = dsize; - audio.body = filename.toStdString(); - audio.url = url.toStdString(); + audio.body = caption.has_value() ? caption.value().toStdString() : filename.toStdString(); + audio.url = url.toStdString(); if (duration > 0) audio.info.duration = duration; @@ -803,6 +820,7 @@ void InputBar::video(const QString &filename, const std::optional &file, const QString &url, + const std::optional &caption, const QString &mime, uint64_t dsize, uint64_t duration, @@ -817,7 +835,7 @@ InputBar::video(const QString &filename, video.info.mimetype = mime.toStdString(); video.info.size = dsize; video.info.blurhash = blurhash.toStdString(); - video.body = filename.toStdString(); + video.body = caption.has_value() ? caption.value().toStdString() : filename.toStdString(); if (duration > 0) video.info.duration = duration; @@ -1282,10 +1300,12 @@ InputBar::finalizeUpload(MediaUpload *upload, const QString &url) auto mimeClass = upload->mimeClass(); auto size = upload->size(); auto encryptedFile = upload->encryptedFile_(); + auto caption = upload->caption(); if (mimeClass == u"image") image(filename, encryptedFile, url, + caption, mime, size, upload->dimensions(), @@ -1295,11 +1315,12 @@ InputBar::finalizeUpload(MediaUpload *upload, const QString &url) upload->thumbnailImg().size(), upload->blurhash()); else if (mimeClass == u"audio") - audio(filename, encryptedFile, url, mime, size, upload->duration()); + audio(filename, encryptedFile, url, caption, mime, size, upload->duration()); else if (mimeClass == u"video") video(filename, encryptedFile, url, + caption, mime, size, upload->duration(), @@ -1310,7 +1331,7 @@ InputBar::finalizeUpload(MediaUpload *upload, const QString &url) upload->thumbnailImg().size(), upload->blurhash()); else - file(filename, encryptedFile, url, mime, size); + file(filename, encryptedFile, url, caption, mime, size); removeRunUpload(upload); } @@ -1405,6 +1426,15 @@ InputBar::acceptUploads() } } +void +InputBar::acceptUploadsWithCaption(QString caption) +{ + for (UploadHandle &upload : unconfirmedUploads) { + upload->caption_ = std::optional(caption); + } + acceptUploads(); +} + void InputBar::declineUploads() { diff --git a/src/timeline/InputBar.h b/src/timeline/InputBar.h index c38de662..a3609642 100644 --- a/src/timeline/InputBar.h +++ b/src/timeline/InputBar.h @@ -86,6 +86,7 @@ public: return MediaType::File; } [[nodiscard]] QString url() const { return url_; } + [[nodiscard]] std::optional caption() const { return caption_; } [[nodiscard]] QString mimetype() const { return mimetype_; } [[nodiscard]] QString mimeClass() const { return mimeClass_; } [[nodiscard]] QString filename() const { return originalFilename_; } @@ -143,6 +144,7 @@ public: QString blurhash_; QString thumbnailUrl_; QString url_; + std::optional caption_; std::optional encryptedFile, thumbnailEncryptedFile; QImage thumbnail_; @@ -236,6 +238,7 @@ public slots: void sticker(QStringList descriptor); void acceptUploads(); + void acceptUploadsWithCaption(QString); void declineUploads(); private slots: @@ -264,6 +267,7 @@ private: void image(const QString &filename, const std::optional &file, const QString &url, + const std::optional &caption, const QString &mime, uint64_t dsize, const QSize &dimensions, @@ -275,17 +279,20 @@ private: void file(const QString &filename, const std::optional &encryptedFile, const QString &url, + const std::optional &caption, const QString &mime, uint64_t dsize); void audio(const QString &filename, const std::optional &file, const QString &url, + const std::optional &caption, const QString &mime, uint64_t dsize, uint64_t duration); void video(const QString &filename, const std::optional &file, const QString &url, + const std::optional &caption, const QString &mime, uint64_t dsize, uint64_t duration,