[ENH] Validate schema in server (#5612)

## Description of changes

_Summarize the changes made by this PR._

- Improvements & Bug fixes
    - N/A
- New functionality
    - Schema validation in server

## Test plan

_How are these changes tested?_

- [ ] Tests pass locally with `pytest` for python, `yarn test` for js, `cargo test` for rust

## Migration plan

_Are there any migrations, or any forwards/backwards compatibility changes needed in order to make sure this change deploys reliably?_

## Observability plan

_What is the plan to instrument and monitor this change?_

## Documentation Changes

_Are all docstrings for user-facing APIs updated if required? Do we need to make documentation changes in the_ [_docs section](https://github.com/chroma-core/chroma/tree/main/docs/docs.trychroma.com)?_

Author: Sicheng-Pan

rust/types/src/api_types.rs

@@ -11,7 +11,7 @@ use crate::plan::PlanToProtoError;
 use crate::plan::SearchPayload;
 use crate::validators::{
     validate_metadata_vec, validate_name, validate_non_empty_collection_update_metadata,
-    validate_optional_metadata, validate_update_metadata_vec,
+    validate_optional_metadata, validate_schema, validate_update_metadata_vec,
 };
 use crate::Collection;
 use crate::CollectionConfigurationToInternalConfigurationError;
@@ -668,6 +668,7 @@ pub struct CreateCollectionRequest {
     #[validate(custom(function = "validate_optional_metadata"))]
     pub metadata: Option<Metadata>,
     pub configuration: Option<InternalCollectionConfiguration>,
+    #[validate(custom(function = "validate_schema"))]
     pub schema: Option<InternalSchema>,
     pub get_or_create: bool,
 }

rust/types/src/validators.rs

@@ -1,6 +1,7 @@
 use crate::{
     operator::{Rank, RankExpr},
-    CollectionMetadataUpdate, Metadata, MetadataValue, UpdateMetadata, UpdateMetadataValue,
+    CollectionMetadataUpdate, InternalSchema, Metadata, MetadataValue, UpdateMetadata,
+    UpdateMetadataValue,
 };
 use regex::Regex;
 use std::collections::HashMap;
@@ -178,6 +179,82 @@ fn validate_rank_expr(expr: &RankExpr) -> Result<(), ValidationError> {
     Ok(())
 }
 
+/// Validate schema
+pub fn validate_schema(schema: &InternalSchema) -> Result<(), ValidationError> {
+    let mut sparse_index_keys = Vec::new();
+    if schema
+        .defaults
+        .float_list
+        .as_ref()
+        .is_some_and(|vt| vt.vector_index.as_ref().is_some_and(|it| it.enabled))
+    {
+        return Err(ValidationError::new("schema").with_message("Vector index cannot be enabled by default. It can only be enabled on #embedding field.".into()));
+    }
+    if schema
+        .defaults
+        .sparse_vector
+        .as_ref()
+        .is_some_and(|vt| vt.sparse_vector_index.as_ref().is_some_and(|it| it.enabled))
+    {
+        return Err(ValidationError::new("schema").with_message("Sparse vector index cannot be enabled by default. Please enable sparse vector index on specific keys. At most one sparse vector index is allowed for the collection.".into()));
+    }
+    if schema
+        .defaults
+        .string
+        .as_ref()
+        .is_some_and(|vt| vt.fts_index.as_ref().is_some_and(|it| it.enabled))
+    {
+        return Err(ValidationError::new("schema").with_message("Full text search / regular expression index cannot be enabled by default. It can only be enabled on #document field.".into()));
+    }
+    for (key, config) in &schema.key_overrides {
+        if let Some(vit) = config
+            .float_list
+            .as_ref()
+            .and_then(|vt| vt.vector_index.as_ref())
+        {
+            // TODO(Sicheng): Schema currently use `$embedding`. This should be updated once schema updates naming
+            if vit.enabled && key != "$embedding" {
+                return Err(ValidationError::new("schema").with_message(
+                    format!("Vector index can only be enabled on $embedding field: {key}").into(),
+                ));
+            }
+            // TODO(Sicheng): Schema currently use `$document`. This should be updated once schema updates naming
+            if vit
+                .config
+                .source_key
+                .as_ref()
+                .is_some_and(|key| key != "$document")
+            {
+                return Err(ValidationError::new("schema")
+                    .with_message("Vector index can only source from $document".into()));
+            }
+        }
+        if config
+            .sparse_vector
+            .as_ref()
+            .is_some_and(|vt| vt.sparse_vector_index.as_ref().is_some_and(|it| it.enabled))
+        {
+            sparse_index_keys.push(key);
+            if sparse_index_keys.len() > 1 {
+                return Err(ValidationError::new("schema").with_message(
+                    format!("At most one sparse vector index is allowed for the collection: {sparse_index_keys:?}")
+                        .into(),
+                ));
+            }
+        }
+        // TODO(Sicheng): Schema currently use `$document`. This should be updated once schema updates naming
+        if config
+            .string
+            .as_ref()
+            .is_some_and(|vt| vt.fts_index.as_ref().is_some_and(|it| it.enabled))
+            && key != "$document"
+        {
+            return Err(ValidationError::new("schema").with_message(format!("Full text search / regular expression index can only be enabled on $document field: {key}").into()));
+        }
+    }
+    Ok(())
+}
+
 #[cfg(test)]
 mod tests {
     use super::*;