feat: real EXIF extraction via adapters-exif crate

- MetadataExtractorPort in domain (bytes → StructuredData)
- adapters-exif: NomExifExtractor using nom-exif, handles EXIF + TrackInfo
- Worker's MetadataExtractorPlugin delegates to port, no longer knows nom-exif
- Filters noisy binary tags (U8Array, Undefined, Unknown)
This commit is contained in:
2026-05-31 20:28:50 +02:00
parent d1c7243f5b
commit 45669ec848
10 changed files with 212 additions and 8 deletions

72
Cargo.lock generated
View File

@@ -17,6 +17,15 @@ dependencies = [
"uuid",
]
[[package]]
name = "adapters-exif"
version = "0.1.0"
dependencies = [
"bytes",
"domain",
"nom-exif",
]
[[package]]
name = "adapters-nats"
version = "0.1.0"
@@ -119,6 +128,15 @@ dependencies = [
"uuid",
]
[[package]]
name = "approx"
version = "0.5.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "cab112f0a86d568ea0e627cc1d6be74a1e9cd55214684db5561995f6dad897c6"
dependencies = [
"num-traits",
]
[[package]]
name = "async-nats"
version = "0.48.0"
@@ -836,6 +854,17 @@ dependencies = [
"version_check",
]
[[package]]
name = "geo-types"
version = "0.7.19"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "94776032c45f950d30a13af6113c2ad5625316c9abfbccee4dd5a6695f8fe0f5"
dependencies = [
"approx",
"num-traits",
"serde",
]
[[package]]
name = "getrandom"
version = "0.2.17"
@@ -1232,6 +1261,16 @@ version = "2.12.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "d98f6fed1fde3f8c21bc40a1abb88dd75e67924f9cffc3ef95607bad8017f8e2"
[[package]]
name = "iso6709parse"
version = "0.1.2"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "b5090db9c6a716d1f4eeb729957e889e9c28156061c825cbccd44950cf0f3c66"
dependencies = [
"geo-types",
"nom",
]
[[package]]
name = "itertools"
version = "0.13.0"
@@ -1385,6 +1424,12 @@ version = "0.3.17"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "6877bb514081ee2a7ff5ef9de3281f14a4dd4bceac4c09388074a6b5df8a139a"
[[package]]
name = "minimal-lexical"
version = "0.2.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "68354c5c6bd36d73ff3feceb05efa59b6acb7626617f4962be322a825e61f79a"
[[package]]
name = "mio"
version = "1.1.1"
@@ -1428,6 +1473,32 @@ dependencies = [
"signatory",
]
[[package]]
name = "nom"
version = "7.1.3"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "d273983c5a657a70a3e8f2a01329822f3b8c8172b73826411a55751e404a0a4a"
dependencies = [
"memchr",
"minimal-lexical",
]
[[package]]
name = "nom-exif"
version = "2.8.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "d759833b65510dc55d774e34b7ef6665ffd293eae44844e189a9da2bea53d47a"
dependencies = [
"bytes",
"chrono",
"iso6709parse",
"nom",
"regex",
"serde",
"thiserror",
"tracing",
]
[[package]]
name = "nu-ansi-term"
version = "0.50.3"
@@ -3563,6 +3634,7 @@ dependencies = [
name = "worker"
version = "0.1.0"
dependencies = [
"adapters-exif",
"adapters-nats",
"adapters-postgres",
"adapters-storage",

View File

@@ -10,6 +10,7 @@ members = [
"crates/adapters/event-payload",
"crates/adapters/event-transport",
"crates/adapters/nats",
"crates/adapters/exif",
"crates/presentation",
"crates/bootstrap",
"crates/worker",
@@ -46,6 +47,7 @@ adapters-storage = { path = "crates/adapters/storage" }
event-payload = { path = "crates/adapters/event-payload" }
event-transport = { path = "crates/adapters/event-transport" }
adapters-nats = { path = "crates/adapters/nats" }
adapters-exif = { path = "crates/adapters/exif" }
async-nats = "0.48"
async-stream = "0.3"
presentation = { path = "crates/presentation" }

View File

@@ -0,0 +1,9 @@
[package]
name = "adapters-exif"
version = "0.1.0"
edition = "2024"
[dependencies]
domain = { workspace = true }
bytes = { workspace = true }
nom-exif = { version = "2.5", features = ["serde"] }

View File

@@ -0,0 +1,80 @@
use bytes::Bytes;
use domain::{
errors::DomainError,
ports::MetadataExtractorPort,
value_objects::{MetadataValue, StructuredData},
};
use nom_exif::{ExifIter, MediaParser, MediaSource, TrackInfo};
use std::io::Cursor;
pub struct NomExifExtractor;
impl MetadataExtractorPort for NomExifExtractor {
fn extract(&self, bytes: &Bytes) -> Result<StructuredData, DomainError> {
if bytes.is_empty() {
return Ok(StructuredData::new());
}
let ms = match MediaSource::seekable(Cursor::new(bytes.as_ref())) {
Ok(ms) => ms,
Err(_) => return Ok(StructuredData::new()),
};
let mut parser = MediaParser::new();
let mut data = StructuredData::new();
if ms.has_exif() {
let iter: ExifIter = match parser.parse(ms) {
Ok(iter) => iter,
Err(_) => return Ok(data),
};
for mut entry in iter {
let tag_name = match entry.tag() {
Some(t) => t.to_string(),
None => continue,
};
if tag_name.starts_with("Unknown(") {
continue;
}
let value = match entry.take_result() {
Ok(v) => v.to_string(),
Err(_) => continue,
};
if is_noisy_value(&value) {
continue;
}
data.insert(tag_name, MetadataValue::String(value));
}
} else {
let track_info = match parser.parse::<_, _, TrackInfo>(ms) {
Ok(info) => info,
Err(_) => return Ok(data),
};
for (key, val) in track_info {
data.insert(
format!("track:{}", key),
MetadataValue::String(val.to_string()),
);
}
}
Ok(data)
}
}
fn is_noisy_value(v: &str) -> bool {
v.starts_with("U16Array")
|| v.starts_with("U32Array")
|| v.starts_with("U8Array")
|| v.starts_with("URationalArray")
|| v.starts_with("Undefined")
}
#[cfg(test)]
mod tests;

View File

@@ -0,0 +1,19 @@
use crate::NomExifExtractor;
use bytes::Bytes;
use domain::ports::MetadataExtractorPort;
#[test]
fn empty_bytes_returns_empty_data() {
let extractor = NomExifExtractor;
let result = extractor.extract(&Bytes::new());
assert!(result.is_ok());
assert!(result.unwrap().is_empty());
}
#[test]
fn garbage_bytes_returns_empty_data() {
let extractor = NomExifExtractor;
let result = extractor.extract(&Bytes::from_static(b"not a real image file"));
assert!(result.is_ok());
assert!(result.unwrap().is_empty());
}

View File

@@ -3,8 +3,9 @@ use super::entities::{
MetadataSource,
};
use crate::common::errors::DomainError;
use crate::common::value_objects::{Checksum, SystemId};
use crate::common::value_objects::{Checksum, StructuredData, SystemId};
use async_trait::async_trait;
use bytes::Bytes;
// --- AssetRepository ---
@@ -74,3 +75,9 @@ pub trait DuplicateRepository: Send + Sync {
async fn find_by_asset(&self, asset_id: &SystemId) -> Result<Vec<DuplicateGroup>, DomainError>;
async fn save(&self, group: &DuplicateGroup) -> Result<(), DomainError>;
}
// --- MetadataExtractorPort ---
pub trait MetadataExtractorPort: Send + Sync {
fn extract(&self, bytes: &Bytes) -> Result<StructuredData, DomainError>;
}

View File

@@ -15,6 +15,7 @@ adapters-postgres = { path = "../adapters/postgres" }
adapters-storage = { workspace = true }
adapters-nats = { workspace = true }
event-transport = { workspace = true }
adapters-exif = { workspace = true }
async-nats = { workspace = true }
futures = { workspace = true }

View File

@@ -1,6 +1,6 @@
use crate::plugin_registry::InMemoryPluginRegistry;
use crate::plugins::{MetadataExtractorPlugin, NoOpPlugin, SidecarSyncPlugin};
use domain::ports::SidecarWriterPort;
use domain::ports::{MetadataExtractorPort, SidecarWriterPort};
use std::sync::Arc;
use super::Repos;
@@ -9,6 +9,7 @@ pub fn build_plugin_registry(
repos: &Repos,
file_storage: Arc<dyn domain::ports::FileStoragePort>,
sidecar_writer: Arc<dyn SidecarWriterPort>,
extractor: Arc<dyn MetadataExtractorPort>,
) -> InMemoryPluginRegistry {
let mut registry = InMemoryPluginRegistry::new();
@@ -17,6 +18,7 @@ pub fn build_plugin_registry(
repos.asset.clone(),
file_storage,
repos.metadata.clone(),
extractor,
)));
let export_handler = Arc::new(application::sidecar::ExportSidecarHandler::new(

View File

@@ -51,7 +51,14 @@ async fn main() -> anyhow::Result<()> {
event_transport::CompositeEventPublisher::new(nats_publisher, event_store),
);
let registry = Arc::new(build_plugin_registry(&repos, file_storage, sidecar_writer));
let extractor: Arc<dyn domain::ports::MetadataExtractorPort> =
Arc::new(adapters_exif::NomExifExtractor);
let registry = Arc::new(build_plugin_registry(
&repos,
file_storage,
sidecar_writer,
extractor,
));
let process_next = Arc::new(build_process_next_handler(
&repos,
registry,

View File

@@ -2,7 +2,10 @@ use async_trait::async_trait;
use domain::{
entities::{AssetMetadata, MetadataSource},
errors::DomainError,
ports::{AssetMetadataRepository, AssetRepository, FileStoragePort, PluginExecutor},
ports::{
AssetMetadataRepository, AssetRepository, FileStoragePort, MetadataExtractorPort,
PluginExecutor,
},
value_objects::{MetadataValue, StructuredData, SystemId},
};
use std::sync::Arc;
@@ -12,6 +15,7 @@ pub struct MetadataExtractorPlugin {
asset_repo: Arc<dyn AssetRepository>,
file_storage: Arc<dyn FileStoragePort>,
metadata_repo: Arc<dyn AssetMetadataRepository>,
extractor: Arc<dyn MetadataExtractorPort>,
}
impl MetadataExtractorPlugin {
@@ -19,11 +23,13 @@ impl MetadataExtractorPlugin {
asset_repo: Arc<dyn AssetRepository>,
file_storage: Arc<dyn FileStoragePort>,
metadata_repo: Arc<dyn AssetMetadataRepository>,
extractor: Arc<dyn MetadataExtractorPort>,
) -> Self {
Self {
asset_repo,
file_storage,
metadata_repo,
extractor,
}
}
}
@@ -52,17 +58,16 @@ impl PluginExecutor for MetadataExtractorPlugin {
let path = &asset.source_reference.relative_path;
let data = self.file_storage.read_file(path).await?;
let file_size = data.len() as i64;
let mut extracted = StructuredData::new();
extracted.insert("file_size_bytes", MetadataValue::Integer(file_size));
let mut extracted = self.extractor.extract(&data)?;
extracted.insert("file_size_bytes", MetadataValue::Integer(data.len() as i64));
extracted.insert("mime_type", MetadataValue::String(asset.mime_type.clone()));
let metadata =
AssetMetadata::new(asset_id, MetadataSource::ExifExtracted, extracted.clone());
self.metadata_repo.save(&metadata).await?;
info!(asset_id = %asset_id, file_size, "extracted basic metadata");
info!(asset_id = %asset_id, tags = extracted.len(), "extracted metadata");
Ok(extracted)
}
}