Add --base-path CLI option to override the URL path in the tilejson (#1205)

Override URL path in the tilejson's `tiles` field when used behind a proxy that is not setting `X-Rewrite-URL` header

Fixes #1185

docs/src/config-file.md

@@ -21,6 +21,10 @@ keep_alive: 75
 # The socket address to bind [default: 0.0.0.0:3000]
 listen_addresses: '0.0.0.0:3000'
 
+
+# Set TileJSON URL path prefix, ignoring X-Rewrite-URL header. Must begin with a `/`
+base_path: /tiles
+
 # Number of web server workers
 worker_processes: 8
 

docs/src/run-with-cli.md

@@ -27,6 +27,8 @@ Options:
 
   -l, --listen-addresses <LISTEN_ADDRESSES>
           The socket address to bind. [DEFAULT: 0.0.0.0:3000]
+      --base-path <BASE_PATH>
+          Set TileJSON URL path prefix, ignoring X-Rewrite-URL header. Must begin with a `/`. Examples: `/`, `/tiles`
           
   -W, --workers <WORKERS>
           Number of web server workers

martin/src/args/mod.rs

@@ -13,5 +13,4 @@ mod root;
 pub use root::{Args, ExtraArgs, MetaArgs};
 
 mod srv;
-pub use srv::PreferredEncoding;
-pub use srv::SrvArgs;
+pub use srv::{PreferredEncoding, SrvArgs};

martin/src/args/srv.rs

@@ -1,7 +1,8 @@
-use crate::srv::{SrvConfig, KEEP_ALIVE_DEFAULT, LISTEN_ADDRESSES_DEFAULT};
 use clap::ValueEnum;
 use serde::{Deserialize, Serialize};
 
+use crate::srv::{SrvConfig, KEEP_ALIVE_DEFAULT, LISTEN_ADDRESSES_DEFAULT};
+
 #[derive(clap::Args, Debug, PartialEq, Default)]
 #[command(about, version)]
 pub struct SrvArgs {
@@ -9,6 +10,9 @@ pub struct SrvArgs {
     pub keep_alive: Option<u64>,
     #[arg(help = format!("The socket address to bind. [DEFAULT: {}]", LISTEN_ADDRESSES_DEFAULT), short, long)]
     pub listen_addresses: Option<String>,
+    /// Set TileJSON URL path prefix, ignoring X-Rewrite-URL header. Must begin with a `/`. Examples: `/`, `/tiles`
+    #[arg(long)]
+    pub base_path: Option<String>,
     /// Number of web server workers
     #[arg(short = 'W', long)]
     pub workers: Option<usize>,
@@ -42,5 +46,8 @@ impl SrvArgs {
         if self.preferred_encoding.is_some() {
             srv_config.preferred_encoding = self.preferred_encoding;
         }
+        if self.base_path.is_some() {
+            srv_config.base_path = self.base_path;
+        }
     }
 }

martin/src/config.rs

@@ -19,7 +19,7 @@ use crate::source::{TileInfoSources, TileSources};
 #[cfg(feature = "sprites")]
 use crate::sprites::{SpriteConfig, SpriteSources};
 use crate::srv::{SrvConfig, RESERVED_KEYWORDS};
-use crate::utils::{CacheValue, MainCache, OptMainCache};
+use crate::utils::{parse_base_path, CacheValue, MainCache, OptMainCache};
 use crate::MartinError::{ConfigLoadError, ConfigParseError, ConfigWriteError, NoSources};
 use crate::{IdResolver, MartinResult, OptOneMany};
 
@@ -71,6 +71,10 @@ impl Config {
         let mut res = UnrecognizedValues::new();
         copy_unrecognized_config(&mut res, "", &self.unrecognized);
 
+        if let Some(path) = &self.srv.base_path {
+            self.srv.base_path = Some(parse_base_path(path)?);
+        }
+
         #[cfg(feature = "postgres")]
         for pg in self.postgres.iter_mut() {
             res.extend(pg.finalize()?);

martin/src/srv/config.rs

@@ -10,6 +10,7 @@ pub const LISTEN_ADDRESSES_DEFAULT: &str = "0.0.0.0:3000";
 pub struct SrvConfig {
     pub keep_alive: Option<u64>,
     pub listen_addresses: Option<String>,
+    pub base_path: Option<String>,
     pub worker_processes: Option<usize>,
     pub preferred_encoding: Option<PreferredEncoding>,
 }
@@ -35,6 +36,7 @@ mod tests {
                 listen_addresses: some("0.0.0.0:3000"),
                 worker_processes: Some(8),
                 preferred_encoding: None,
+                base_path: None,
             }
         );
         assert_eq!(
@@ -50,6 +52,7 @@ mod tests {
                 listen_addresses: some("0.0.0.0:3000"),
                 worker_processes: Some(8),
                 preferred_encoding: Some(PreferredEncoding::Brotli),
+                base_path: None
             }
         );
         assert_eq!(
@@ -65,6 +68,7 @@ mod tests {
                 listen_addresses: some("0.0.0.0:3000"),
                 worker_processes: Some(8),
                 preferred_encoding: Some(PreferredEncoding::Brotli),
+                base_path: None,
             }
         );
     }

martin/src/srv/tiles_info.rs

@@ -9,6 +9,7 @@ use serde::Deserialize;
 use tilejson::{tilejson, TileJSON};
 
 use crate::source::{Source, TileSources};
+use crate::srv::SrvConfig;
 
 #[derive(Deserialize)]
 pub struct SourceIDsRequest {
@@ -26,17 +27,19 @@ async fn get_source_info(
     req: HttpRequest,
     path: Path<SourceIDsRequest>,
     sources: Data<TileSources>,
+    srv_config: Data<SrvConfig>,
 ) -> ActixResult<HttpResponse> {
     let sources = sources.get_sources(&path.source_ids, None)?.0;
 
-    // Get `X-REWRITE-URL` header value, and extract its `path` component.
-    // If the header is not present or cannot be parsed as a URL, return the request path.
-    let tiles_path = req
-        .headers()
+    let tiles_path = if let Some(base_path) = &srv_config.base_path {
+        format!("{base_path}/{}", path.source_ids)
+    } else {
+        req.headers()
             .get("x-rewrite-url")
             .and_then(|v| v.to_str().ok())
             .and_then(|v| v.parse::<Uri>().ok())
-        .map_or_else(|| req.path().to_owned(), |v| v.path().to_owned());
+            .map_or_else(|| req.path().to_owned(), |v| v.path().to_owned())
+    };
 
     let query_string = req.query_string();
     let path_and_query = if query_string.is_empty() {
@@ -155,7 +158,7 @@ pub fn merge_tilejson(sources: &[&dyn Source], tiles_url: String) -> TileJSON {
 pub mod tests {
     use std::collections::BTreeMap;
 
-    use tilejson::{tilejson, Bounds, VectorLayer};
+    use tilejson::{Bounds, VectorLayer};
 
     use super::*;
     use crate::srv::server::tests::TestSource;

martin/src/utils/error.rs

@@ -43,6 +43,9 @@ pub enum MartinError {
     #[error("Unable to bind to {1}: {0}")]
     BindingError(io::Error, String),
 
+    #[error("Base path must be a valid URL path, and must begin with a '/' symbol, but is '{0}'")]
+    BasePathError(String),
+
     #[error("Unable to load config file {}: {0}", .1.display())]
     ConfigLoadError(io::Error, PathBuf),
 

martin/src/utils/utilities.rs

@@ -1,8 +1,12 @@
 use std::io::{Read as _, Write as _};
 
+use actix_web::http::Uri;
 use flate2::read::GzDecoder;
 use flate2::write::GzEncoder;
 
+use crate::MartinError::BasePathError;
+use crate::MartinResult;
+
 pub fn decode_gzip(data: &[u8]) -> Result<Vec<u8>, std::io::Error> {
     let mut decoder = GzDecoder::new(data);
     let mut decompressed = Vec::new();
@@ -28,3 +32,34 @@ pub fn encode_brotli(data: &[u8]) -> Result<Vec<u8>, std::io::Error> {
     encoder.write_all(data)?;
     Ok(encoder.into_inner())
 }
+
+pub fn parse_base_path(path: &str) -> MartinResult<String> {
+    if !path.starts_with('/') {
+        return Err(BasePathError(path.to_string()));
+    }
+    if let Ok(uri) = path.parse::<Uri>() {
+        return Ok(uri.path().trim_end_matches('/').to_string());
+    }
+    Err(BasePathError(path.to_string()))
+}
+
+#[cfg(test)]
+pub mod tests {
+    use crate::utils::parse_base_path;
+    #[test]
+    fn test_parse_base_path() {
+        for (path, expected) in [
+            ("/", Some("")),
+            ("//", Some("")),
+            ("/foo/bar", Some("/foo/bar")),
+            ("/foo/bar/", Some("/foo/bar")),
+            ("", None),
+            ("foo/bar", None),
+        ] {
+            match expected {
+                Some(v) => assert_eq!(v, parse_base_path(path).unwrap()),
+                None => assert!(parse_base_path(path).is_err()),
+            }
+        }
+    }
+}