BLOG-78 Backend image upload and download #84

squid · 2025-07-27T12:26:50+08:00

squid commented

2025-07-27 12:26:50 +08:00

Description

Add some endpoints about image:
- POST /image/upload
- GET /image/{id}

Note

Since there isn't identity authentication, the /image endpoints should be restricted to private network in nginx.

Note

Volume for backend should be configured in pod.yaml.

Package Changes

actix-multipart = "0.7.2"

Screenshots

No response

Reference

Resolves #78

Checklist

A milestone is set
The related issuse has been linked to this branch

### Description - Add some endpoints about image: - POST `/image/upload` - GET `/image/{id}` > [!NOTE] > Since there isn't identity authentication, the `/image` endpoints should be restricted to private network in nginx. > [!NOTE] > Volume for backend should be configured in `pod.yaml`. ### Package Changes ```toml actix-multipart = "0.7.2" ``` ### Screenshots _No response_ ### Reference Resolves #78 ### Checklist - [x] A milestone is set - [x] The related issuse has been linked to this branch

squid added this to the 0.3 milestone 2025-07-27 12:26:50 +08:00

squid added 4 commits 2025-07-27 12:26:51 +08:00

BLOG-78 refactor: rename db mapper

Frontend CI / build (push) Successful in 1m5s

Details

7b431e3fc8

BLOG-78 feat: implement image upload and retrieval feature

Frontend CI / build (push) Successful in 1m9s

Details

d58a7d1a37

This commit introduces the capability to upload and serve images.

A new `image` feature module has been added to the backend, following the existing clean architecture pattern.

- Implements `POST /image/upload` for uploading image files.
- Implements `GET /image/{id}` for retrieving an image by its ID.
- Adds a new `image` table to the database to store image metadata.
- Image data is stored in the file system. The path can be configured via the `STORAGE_PATH` environment variable.

BLOG-78 feat: add MIME type validation for image uploads

Frontend CI / build (push) Successful in 1m8s

Details

c639a85ff3

BLOG-78 feat: add volume and storage path for static files in Dockerfile

Frontend CI / build (push) Successful in 1m5s

Details

PR Title Check / pr-title-check (pull_request) Successful in 13s

Details

Auto Comment On PR / add_improve_comment (pull_request) Successful in 13s

Details

5a30642b8f

bot commented

2025-07-27 12:42:01 +08:00

/improve

bot commented

2025-07-27 12:42:07 +08:00

PR Code Suggestions ✨

Category	Suggestion	Impact
General	Avoid unnecessary data cloning The `to_entity` method currently clones the `data` vector. If `ImageRequestDto` is a transient object used solely to create an `Image` entity, it can take ownership of the `data` to avoid an unnecessary deep copy, which can be inefficient for large images. Change the method to consume `self`. backend/feature/image/src/adapter/delivery/image_request_dto.rs [9-14] `-pub fn to_entity(&self) -> Image { +pub fn to_entity(self) -> Image { Image { id: None, - mime_type: self.mime_type.clone(), - data: self.data.clone(), + mime_type: self.mime_type, + data: self.data, } }` Suggestion importance[1-10]: 9 __ Why: This is a significant optimization. By consuming `self` in `to_entity`, the method avoids unnecessary deep cloning of the `data` vector, which can be large, leading to better performance and more idiomatic Rust.	High
General	Construct complete objects directly The current implementation constructs an `Image` entity in two steps: first, an incomplete `Image` with empty data from `ImageDbMapper::to_entity`, then mutating it with data from storage. It's more robust and idiomatic to construct the `Image` object fully in one step, combining metadata from `image_mapper` and data from `image_storage`. This also allows removing the problematic `to_entity` method from `ImageDbMapper`. backend/feature/image/src/adapter/gateway/image_repository_impl.rs [44-47] `let image_mapper = self.image_db_service.get_image_info_by_id(id).await?; -let mut image = image_mapper.to_entity(); -image.data = self.image_storage.read_data(id)?; -Ok(image) +let data = self.image_storage.read_data(id)?; +Ok(Image { + id: image_mapper.id, + mime_type: image_mapper.mime_type, + data, +})` Suggestion importance[1-10]: 8 __ Why: This suggestion improves code clarity and robustness by constructing the `Image` object in a single step, avoiding an intermediate, incomplete state and an unnecessary `Vec::new()` allocation. It's a good idiomatic Rust practice.	Medium

## PR Code Suggestions ✨  <table><thead><tr><td><strong>Category</strong></td><td align=left><strong>Suggestion                                                                                                                                    </strong></td><td align=center><strong>Impact</strong></td></tr><tbody><tr><td rowspan=2>General</td> <td> <details><summary>Avoid unnecessary data cloning</summary> ___ **The <code>to_entity</code> method currently clones the <code>data</code> vector. If <code>ImageRequestDto</code> is a <br>transient object used solely to create an <code>Image</code> entity, it can take ownership of the <br><code>data</code> to avoid an unnecessary deep copy, which can be inefficient for large images. <br>Change the method to consume <code>self</code>.** [backend/feature/image/src/adapter/delivery/image_request_dto.rs [9-14]](https://git.squidspirit.com/squid/blog/src/branch/BLOG-78_image_management/backend/feature/image/src/adapter/delivery/image_request_dto.rs#L9-L14) ```diff -pub fn to_entity(&self) -> Image { +pub fn to_entity(self) -> Image { Image { id: None, - mime_type: self.mime_type.clone(), - data: self.data.clone(), + mime_type: self.mime_type, + data: self.data, } } ``` <details><summary>Suggestion importance[1-10]: 9</summary> __ Why: This is a significant optimization. By consuming `self` in `to_entity`, the method avoids unnecessary deep cloning of the `data` vector, which can be large, leading to better performance and more idiomatic Rust. </details></details></td><td align=center>High </td></tr><tr><td> <details><summary>Construct complete objects directly</summary> ___ **The current implementation constructs an <code>Image</code> entity in two steps: first, an <br>incomplete <code>Image</code> with empty data from <code>ImageDbMapper::to_entity</code>, then mutating it <br>with data from storage. It's more robust and idiomatic to construct the <code>Image</code> object <br>fully in one step, combining metadata from <code>image_mapper</code> and data from <code>image_storage</code>. <br>This also allows removing the problematic <code>to_entity</code> method from <code>ImageDbMapper</code>.** [backend/feature/image/src/adapter/gateway/image_repository_impl.rs [44-47]](https://git.squidspirit.com/squid/blog/src/branch/BLOG-78_image_management/backend/feature/image/src/adapter/gateway/image_repository_impl.rs#L44-L47) ```diff let image_mapper = self.image_db_service.get_image_info_by_id(id).await?; -let mut image = image_mapper.to_entity(); -image.data = self.image_storage.read_data(id)?; -Ok(image) +let data = self.image_storage.read_data(id)?; +Ok(Image { + id: image_mapper.id, + mime_type: image_mapper.mime_type, + data, +}) ``` <details><summary>Suggestion importance[1-10]: 8</summary> __ Why: This suggestion improves code clarity and robustness by constructing the `Image` object in a single step, avoiding an intermediate, incomplete state and an unnecessary `Vec::new()` allocation. It's a good idiomatic Rust practice. </details></details></td><td align=center>Medium </td></tr></tr></tbody></table>

squid added 1 commit 2025-07-27 12:52:55 +08:00

BLOG-78 refactor: simplify get_image_by_id implementation in ImageRepositoryImpl

PR Title Check / pr-title-check (pull_request) Successful in 13s

Details

Frontend CI / build (push) Successful in 1m8s

Details

aaf43f21bd

squid commented

2025-07-27 12:54:23 +08:00

Construct complete objects directly

The current implementation constructs an Image entity in two steps: first, an
incomplete Image with empty data from ImageDbMapper::to_entity, then mutating it
with data from storage. It's more robust and idiomatic to construct the Image object
fully in one step, combining metadata from image_mapper and data from image_storage.
This also allows removing the problematic to_entity method from ImageDbMapper.

backend/feature/image/src/adapter/gateway/image_repository_impl.rs [44-47]
 let image_mapper = self.image_db_service.get_image_info_by_id(id).await?;
-let mut image = image_mapper.to_entity();
-image.data = self.image_storage.read_data(id)?;
-Ok(image)
+let data = self.image_storage.read_data(id)?;
+Ok(Image {
+    id: image_mapper.id,
+    mime_type: image_mapper.mime_type,
+    data,
+})

Addressed in aaf43f21bd

> Construct complete objects directly > > The current implementation constructs an <code>Image</code> entity in two steps: first, an <br>incomplete <code>Image</code> with empty data from <code>ImageDbMapper::to_entity</code>, then mutating it <br>with data from storage. It's more robust and idiomatic to construct the <code>Image</code> object <br>fully in one step, combining metadata from <code>image_mapper</code> and data from <code>image_storage</code>. <br>This also allows removing the problematic <code>to_entity</code> method from <code>ImageDbMapper</code>. > > [backend/feature/image/src/adapter/gateway/image_repository_impl.rs [44-47]](https://git.squidspirit.com/squid/blog/src/branch/BLOG-78_image_management/backend/feature/image/src/adapter/gateway/image_repository_impl.rs#L44-L47) > > ```diff > let image_mapper = self.image_db_service.get_image_info_by_id(id).await?; > -let mut image = image_mapper.to_entity(); > -image.data = self.image_storage.read_data(id)?; > -Ok(image) > +let data = self.image_storage.read_data(id)?; > +Ok(Image { > + id: image_mapper.id, > + mime_type: image_mapper.mime_type, > + data, > +}) > ``` Addressed in aaf43f21bd12f4ca0dcea67288625d95fb56a189

squid added 1 commit 2025-07-27 13:04:28 +08:00

BLOG-78 refactor: rename to_entity methods to into_entity for consistency

PR Title Check / pr-title-check (pull_request) Successful in 16s

Details

Frontend CI / build (push) Successful in 1m5s

Details

f5c12ce56d

squid commented

2025-07-27 13:07:39 +08:00

Avoid unnecessary data cloning

The to_entity method currently clones the data vector. If ImageRequestDto is a
transient object used solely to create an Image entity, it can take ownership of the
data to avoid an unnecessary deep copy, which can be inefficient for large images.
Change the method to consume self.

backend/feature/image/src/adapter/delivery/image_request_dto.rs [9-14]
-pub fn to_entity(&self) -> Image {
+pub fn to_entity(self) -> Image {
     Image {
         id: None,
-        mime_type: self.mime_type.clone(),
-        data: self.data.clone(),
+        mime_type: self.mime_type,
+        data: self.data,
     }
 }

Addressed in f5c12ce56d

> Avoid unnecessary data cloning > > > The <code>to_entity</code> method currently clones the <code>data</code> vector. If <code>ImageRequestDto</code> is a <br>transient object used solely to create an <code>Image</code> entity, it can take ownership of the <br><code>data</code> to avoid an unnecessary deep copy, which can be inefficient for large images. <br>Change the method to consume <code>self</code>. > > [backend/feature/image/src/adapter/delivery/image_request_dto.rs [9-14]](https://git.squidspirit.com/squid/blog/src/branch/BLOG-78_image_management/backend/feature/image/src/adapter/delivery/image_request_dto.rs#L9-L14) > > ```diff > -pub fn to_entity(&self) -> Image { > +pub fn to_entity(self) -> Image { > Image { > id: None, > - mime_type: self.mime_type.clone(), > - data: self.data.clone(), > + mime_type: self.mime_type, > + data: self.data, > } > } > ``` Addressed in f5c12ce56d6f9390e544869ffb2f0b68d817637d

squid merged commit ab3050db69 into main

2025-07-27 13:10:47 +08:00

squid deleted branch BLOG-78_image_management

2025-07-27 13:10:47 +08:00

squid referenced this issue from a commit

2025-07-27 13:10:47 +08:00

BLOG-78 Backend image upload and download (#84)

squid referenced this pull request

2025-07-30 03:48:46 +08:00

Image can only be uploaded by logged in user #86

Sign in to join this conversation.