Keywords based on harvested files

.github/ISSUE_TEMPLATE/add-plugin-to-marketplace.yml

@@ -48,6 +48,13 @@ body:
       multiple: true
       options: ["harvest", "process", "curate", "deposit", "postprocess"]
 
+  - id: "harvested-files"
+    type: "input"
+    attributes:
+      label: "Harvested files"
+      description: "The types of files your plugin harvests (if it is a harvest plugin)"
+      placeholder: "foobar, foobar.yml, foobar.json"
+
   - id: "repository-url"
     type: "input"
     attributes:

docs/source/_ext/plugin_markup.py

@@ -3,6 +3,7 @@
 # SPDX-FileContributor: David Pape
 
 import json
+import re
 from pathlib import Path
 from typing import Any, Dict
 
@@ -30,6 +31,19 @@ def log_message(text: str, text2: str = None) -> None:
     logger.info(message)
 
 
+def keywordify(text: str) -> str:
+    """Make keyword-friendly text.
+
+    The result will only contain lowercase a through z and hyphens, e.g.:
+
+    * CITATION.cff → citation-cff
+    * codemeta.json → codemeta-json
+    * LICENSE → license
+    """
+    text = text.casefold()
+    return re.sub(r"[^a-z]", "-", text)
+
+
 def plugin_to_schema_org(plugin: Dict[str, Any]) -> SchemaOrgSoftwarePublication:
     """Convert plugin metadata from the used JSON format to Schema.org.
 
@@ -43,6 +57,11 @@ def plugin_to_schema_org(plugin: Dict[str, Any]) -> SchemaOrgSoftwarePublication
     """
     keywords = [f"hermes-step-{step}" for step in plugin.get("steps", [])]
 
+    if "harvest" in plugin.get("steps", []) and (
+        harvested_files := plugin.get("harvested_files", [])
+    ):
+        keywords += [f"hermes-harvest-{keywordify(file)}" for file in harvested_files]
+
     return SchemaOrgSoftwarePublication(
         name=plugin.get("name"),
         url=plugin.get("repository_url"),

docs/source/plugins-schema.json

@@ -6,6 +6,7 @@
   "minItems": 1,
   "items": {
     "type": "object",
+    "additionalProperties": false,
     "properties": {
       "name": {
         "type": "string",
@@ -34,6 +35,14 @@
           ]
         }
       },
+      "harvested_files": {
+        "type": "array",
+        "description": "Names of files harvested by a harvest plugin",
+        "items": {
+          "type": "string",
+          "uniqueItems": true
+        }
+      },
       "repository_url": {
         "type": "string",
         "description": "Link to the repository where users can find and inspect the source code of the plugin",

docs/source/plugins.json

@@ -4,13 +4,15 @@
         "description": "Harvest plugin for CITATION.cff files.",
         "author": "Hermes team",
         "steps": ["harvest"],
+        "harvested_files": ["CITATION.cff"],
         "builtin": true
     },
     {
         "name": "codemeta.json Harvester",
         "description": "Harvest plugin for codemeta.json files.",
         "author": "Hermes team",
         "steps": ["harvest"],
+        "harvested_files": ["codemeta.json"],
         "builtin": true
     },
     {
@@ -25,6 +27,7 @@
         "description": "Harvest plugin for .toml files.",
         "author": "Hermes team",
         "steps": ["harvest"],
+        "harvested_files": ["pyproject.toml"],
         "repository_url": "https://github.com/softwarepub/hermes-plugin-python",
         "pypi_url": "https://pypi.org/project/hermes-plugin-python/"
     },