googleapis/mcp-toolbox · error

invalid path: expected 'projects', got %q

Error message

invalid path: expected 'projects', got %q

What it means

The page URL path is validated segment-by-segment against the canonical Healthcare API layout (v1/projects/<project>/locations/<location>/datasets/<ds>/fhirStores/<store>/...). This error means segment[1] is not the literal string "projects", so the URL does not follow the required structure and is rejected before any request is made.

Source

Thrown at internal/sources/cloudhealthcare/cloud_healthcare.go:359

	}

	// Clean and split path
	cleanPath := path.Clean(parsed.Path)
	// Truncate leading and trailing slashes for easier splitting
	trimmed := strings.Trim(cleanPath, "/")
	parts := strings.Split(trimmed, "/")

	// Page URL format Reference: https://docs.cloud.google.com/healthcare-api/docs/how-tos/fhir-search#using_the_search_method_with_get
	if len(parts) < 10 {
		return "", fmt.Errorf("invalid FHIR URL path structure: path too short")
	}

	if !isValidAPIVersion(parts[0]) {
		return "", fmt.Errorf("invalid API version prefix: %q", parts[0])
	}

	if parts[1] != "projects" {
		return "", fmt.Errorf("invalid path: expected 'projects', got %q", parts[1])
	}
	if parts[2] != s.Project() {
		return "", fmt.Errorf("invalid project %q: must match source project %q", parts[2], s.Project())
	}
	if parts[3] != "locations" {
		return "", fmt.Errorf("invalid path: expected 'locations', got %q", parts[3])
	}
	if parts[4] != s.Region() {
		return "", fmt.Errorf("invalid location/region %q: must match source region %q", parts[4], s.Region())
	}
	if parts[5] != "datasets" {
		return "", fmt.Errorf("invalid path: expected 'datasets', got %q", parts[5])
	}
	if parts[6] != s.DatasetID() {
		return "", fmt.Errorf("invalid dataset ID %q: must match source dataset %q", parts[6], s.DatasetID())
	}
	if parts[7] != "fhirStores" {
		return "", fmt.Errorf("invalid path: expected 'fhirStores', got %q", parts[7])

View on GitHub (pinned to 8cc6e09de2)

Solutions

1. Use the exact links.next URL from the previous Google response — it always has "projects" as the second segment
2. Remove any proxy path prefixes from the page URL or decode %2F-encoded slashes causing misalignment
3. Fix casing: segments are lowercase and case-sensitive ("projects", not "Projects")
4. Count the segments: v1(0)/projects(1)/<project>(2)/locations(3)/...

Example fix

// before (proxy prefix shifted segments)
page := "https://healthcare.googleapis.com/v1/api/projects/p/locations/l/..."
// after (canonical Google path)
page := "https://healthcare.googleapis.com/v1/projects/p/locations/l/datasets/d/fhirStores/f/fhir/Patient?pageToken=x"

function hasCanonicalSegments(u) { const p = new URL(u).pathname.split('/').filter(Boolean); return p[0]?.startsWith('v') && p[1] === 'projects' && p[3] === 'locations'; }

When it happens

Trigger: A page URL whose second path segment differs from "projects" — e.g. a relative path that lost its first segments, a URL with the version segment doubled, or a proxy path prefix like /api/projects/... shifting all segments.

Common situations: Reverse proxy base-path prefixes changing segment order; LLM-written URLs with typos ("project" singular, "Projects" capitalized — comparison is case-sensitive); URL-encoding of the segment.

Understand the failure class

Background: "Invalid URL" errors: why new URL(), URI.parse, and reqwest::Url reject your string — missing scheme, whitespace, and bad path format — this error's family across 39 libraries.

Related errors

• invalid FHIR URL path structure: path too short
• invalid API version prefix: %q
• invalid page URL: %w
• no document found
• unable to iterate through query results: %w

AI-assisted analysis of googleapis/mcp-toolbox@8cc6e09de2 (2026-09-05). Data as JSON: /api/errors/bd814b0bac525677. Report an issue: GitHub.