GDCC/7715 Signed Urls for external tools

doc/release-notes/7715-signed-urls-for-external-tools.md

@@ -0,0 +1,3 @@
+# Improved Security for External Tools
+
+This release adds support for configuring external tools to use signed URLs to access the Dataverse API. This eliminates the need for tools to have access to the user's apiToken in order to access draft or restricted datasets and datafiles. Signed URLS can be transferred via POST or via a callback when triggering a tool via GET.

doc/sphinx-guides/source/_static/admin/dataverse-external-tools.tsv

@@ -1,5 +1,5 @@
 Tool	Type	Scope	Description
 Data Explorer	explore	file	A GUI which lists the variables in a tabular data file allowing searching, charting and cross tabulation analysis. See the README.md file at https://github.com/scholarsportal/dataverse-data-explorer-v2 for the instructions on adding Data Explorer to your Dataverse. 
 Whole Tale	explore	dataset	A platform for the creation of reproducible research packages that allows users to launch containerized interactive analysis environments based on popular tools such as Jupyter and RStudio. Using this integration, Dataverse users can launch Jupyter and RStudio environments to analyze published datasets. For more information, see the `Whole Tale User Guide <https://wholetale.readthedocs.io/en/stable/users_guide/integration.html>`_.
-File Previewers	explore	file	A set of tools that display the content of files - including audio, html, `Hypothes.is <https://hypothes.is/>`_ annotations, images, PDF, text, video, tabular data, spreadsheets, and GeoJSON - allowing them to be viewed without downloading. The previewers can be run directly from github.io, so the only required step is using the Dataverse API to register the ones you want to use. Documentation, including how to optionally brand the previewers, and an invitation to contribute through github are in the README.md file. Initial development was led by the Qualitative Data Repository and the spreasdheet previewer was added by the Social Sciences and Humanities Open Cloud (SSHOC) project. https://github.com/gdcc/dataverse-previewers
+File Previewers	explore	file	A set of tools that display the content of files - including audio, html, `Hypothes.is <https://hypothes.is/>`_ annotations, images, PDF, text, video, tabular data, spreadsheets, GeoJSON, and ZipFiles - allowing them to be viewed without downloading the file. The previewers can be run directly from github.io, so the only required step is using the Dataverse API to register the ones you want to use. Documentation, including how to optionally brand the previewers, and an invitation to contribute through github are in the README.md file. Initial development was led by the Qualitative Data Repository and the spreasdheet previewer was added by the Social Sciences and Humanities Open Cloud (SSHOC) project. https://github.com/gdcc/dataverse-previewers
 Data Curation Tool	configure	file	A GUI for curating data by adding labels, groups, weights and other details to assist with informed reuse. See the README.md file at https://github.com/scholarsportal/Dataverse-Data-Curation-Tool for the installation instructions.

doc/sphinx-guides/source/_static/installation/files/root/external-tools/dynamicDatasetTool.json

@@ -12,8 +12,16 @@
         "PID": "{datasetPid}"
       },
       {
-        "apiToken": "{apiToken}"
+        "locale":"{localeCode}"
       }
-    ]
+    ],
+    "allowedApiCalls": [
+        {
+          "name":"retrieveDatasetJson",
+          "httpMethod":"GET",
+          "urlTemplate":"/api/v1/datasets/{datasetId}",
+          "timeOut":10
+        }
+      ]
   }
 }

doc/sphinx-guides/source/_static/installation/files/root/external-tools/fabulousFileTool.json

@@ -1,6 +1,6 @@
 {
   "displayName": "Fabulous File Tool",
-  "description": "Fabulous Fun for Files!",
+  "description": "A non-existent tool that is fabulous fun for files!",
   "toolName": "fabulous",
   "scope": "file",
   "types": [
@@ -9,13 +9,25 @@
   ],
   "toolUrl": "https://fabulousfiletool.com",
   "contentType": "text/tab-separated-values",
+  "httpMethod":"GET",
   "toolParameters": {
     "queryParameters": [
       {
         "fileid": "{fileId}"
       },
       {
-        "key": "{apiToken}"
+        "datasetPid": "{datasetPid}"
+      },
+      {
+        "locale":"{localeCode}"
+      }
+    ],
+    "allowedApiCalls": [
+      {
+        "name":"retrieveDataFile",
+        "httpMethod":"GET",
+        "urlTemplate":"/api/v1/access/datafile/{fileId}",
+        "timeOut":270
       }
     ]
   }

doc/sphinx-guides/source/api/external-tools.rst

@@ -92,7 +92,9 @@ Terminology
 
     contentType                  File level tools operate on a specific **file type** (content type or MIME type such as "application/pdf") and this must be specified. Dataset level tools do not use contentType.
 
-    toolParameters               **Query parameters** are supported and described below.
+    toolParameters               **httpMethod**, **queryParameters**, and **allowedApiCalls** are supported and described below.
+
+    httpMethod                   Either ``GET`` or ``POST``.
 
     queryParameters              **Key/value combinations** that can be appended to the toolUrl. For example, once substitution takes place (described below) the user may be redirected to ``https://fabulousfiletool.com?fileId=42&siteUrl=http://demo.dataverse.org``.
 
@@ -102,6 +104,16 @@ Terminology
 
     reserved words               A **set of strings surrounded by curly braces** such as ``{fileId}`` or ``{datasetId}`` that will be inserted into query parameters. See the table below for a complete list.
 
+    allowedApiCalls              An array of objects defining callbacks the tool is allowed to make to the Dataverse API. If the dataset or file being accessed is not public, the callback URLs will be signed to allow the tool access for a defined time.
+
+    allowedApiCalls name         A name the tool will use to identify this callback URL such as ``retrieveDataFile``.
+
+    allowedApiCalls urlTemplate  The relative URL for the callback using reserved words to indicate where values should by dynamically substituted such as ``/api/v1/datasets/{datasetId}``.
+
+    allowedApiCalls httpMethod   Which HTTP method the specified callback uses such as ``GET`` or ``POST``.
+
+    allowedApiCalls timeOut      For non-public datasets and datafiles, how many minutes the signed URLs given to the tool should be valid for. Must be an integer.
+
     toolName                     A **name** of an external tool that is used to differentiate between external tools and also used in bundle.properties for localization in the Dataverse installation web interface. For example, the toolName for Data Explorer is ``explorer``. For the Data Curation Tool the toolName is ``dct``. This is an optional parameter in the manifest JSON file.   
     ===========================  ==========
 
@@ -131,6 +143,25 @@ Reserved Words
     ``{localeCode}``             optional    The code for the language ("en" for English, "fr" for French, etc.) that user has selected from the language toggle in a Dataverse installation. See also :ref:`i18n`.
     ===========================  ==========  ===========
 
+.. _api-exttools-auth:
+
+Authorization Options
++++++++++++++++++++++
+
+When called for datasets or data files that are not public (i.e. in a draft dataset or for a restricted file), external tools are allowed access via the user's credentials. This is accomplished by one of two mechanisms:
+
+* Signed URLs (more secure, recommended)
+
+  - Configured via the ``allowedApiCalls`` section of the manifest. The tool will be provided with signed URLs allowing the specified access to the given dataset or datafile for the specified amount of time. The tool will not be able to access any other datasets or files the user may have access to and will not be able to make calls other than those specified.
+  - For tools invoked via a GET call, Dataverse will include a callback query parameter with a Base64 encoded value. The decoded value is a signed URL that can be called to retrieve a JSON response containing all of the queryParameters and allowedApiCalls specified in the manfiest.
+  - For tools invoked via POST, Dataverse will send a JSON body including the requested queryParameters and allowedApiCalls. Dataverse expects the response to the POST to indicate a redirect which Dataverse will use to open the tool.
+
+* API Token (deprecated, less secure, not recommended)
+
+  - Configured via the ``queryParameters`` by including an ``{apiToken}`` value. When this is present Dataverse will send the user's apiToken to the tool. With the user's API token, the tool can perform any action via the Dataverse API that the user could. External tools configured via this method should be assessed for their trustworthiness.
+  - For tools invoked via GET, this will be done via a query parameter in the request URL which could be cached in the browser's history. Dataverse expects the response to the POST to indicate a redirect which Dataverse will use to open the tool.
+  - For tools invoked via POST, Dataverse will send a JSON body including the apiToken.
+
 Internationalization of Your External Tool
 ++++++++++++++++++++++++++++++++++++++++++
 

doc/sphinx-guides/source/api/native-api.rst

@@ -2029,6 +2029,24 @@ Archiving is an optional feature that may be configured for a Dataverse installa
 
   curl -H "X-Dataverse-key: $API_TOKEN" -X DELETE "$SERVER_URL/api/datasets/:persistentId/$VERSION/archivalStatus?persistentId=$PERSISTENT_IDENTIFIER"
 
+Get External Tool Parameters
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+This API call is intended as a callback that can be used by :doc:`/installation/external-tools` to retrieve signed Urls  necessary for their interaction with Dataverse.
+It can be called directly as well.
+
+The response is a JSON object described in the :doc:`/api/external-tools` section of the API guide.
+
+.. code-block:: bash
+
+  export API_TOKEN=xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx
+  export SERVER_URL=https://demo.dataverse.org
+  export PERSISTENT_IDENTIFIER=doi:10.5072/FK2/7U7YBV
+  export VERSION=1.0
+  export TOOL_ID=1
+
+
+  curl -H "X-Dataverse-key: $API_TOKEN" -H "Accept:application/json" "$SERVER_URL/api/datasets/:persistentId/versions/$VERSION/toolparams/$TOOL_ID?persistentId=$PERSISTENT_IDENTIFIER"
 
 Files
 -----
@@ -2689,6 +2707,24 @@ Note the optional "limit" parameter. Without it, the API will attempt to populat
 
 By default, the admin API calls are blocked and can only be called from localhost. See more details in :ref:`:BlockedApiEndpoints <:BlockedApiEndpoints>` and :ref:`:BlockedApiPolicy <:BlockedApiPolicy>` settings in :doc:`/installation/config`.
 
+Get External Tool Parameters
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+This API call is intended as a callback that can be used by :doc:`/installation/external-tools` to retrieve signed Urls  necessary for their interaction with Dataverse.
+It can be called directly as well.
+
+The response is a JSON object described in the :doc:`/api/external-tools` section of the API guide.
+
+.. code-block:: bash
+
+  export API_TOKEN=xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx
+  export SERVER_URL=https://demo.dataverse.org
+  export FILE_ID=3
+  export FILEMETADATA_ID=1
+  export TOOL_ID=1
+
+  curl -H "X-Dataverse-key: $API_TOKEN" -H "Accept:application/json" "$SERVER_URL/api/files/$FILE_ID/metadata/$FILEMETADATA_ID/toolparams/$TOOL_ID
+
 Users Token Management
 ----------------------
 
@@ -4070,6 +4106,33 @@ The fully expanded example above (without environment variables) looks like this
 .. code-block:: bash
 
   curl -X DELETE https://demo.dataverse.org/api/admin/template/24
+
+.. _api-native-signed-url:
 
-
-
+Request Signed URL
+~~~~~~~~~~~~~~~~~~
+
+Dataverse has the ability to create signed URLs for it's API calls.
+A signature, which is valid only for the specific API call and only for a specified duration, allows the call to proceed with the authentication of the specified user.
+It is intended as an alternative to the use of an API key (which is valid for a long time period and can be used with any API call).
+Signed URLs were developed to support External Tools but may be useful in other scenarios where Dataverse or a third-party tool needs to delegate limited access to another user or tool. 
+This API call allows a Dataverse superUser to generate a signed URL for such scenarios.
+The JSON input parameter required is an object with the following keys:
+
+- ``url`` - the exact URL to sign, including api version number and all query parameters
+- ``timeOut`` - how long in minutes the signature should be valid for, default is 10 minutes
+- ``httpMethod`` - which HTTP method is required, default is GET
+- ``user`` - the user identifier for the account associated with this signature, the default is the superuser making the call. The API call will succeed/fail based on whether the specified user has the required permissions. 
+
+A curl example using allowing access to a dataset's metadata
+
+.. code-block:: bash
+
+  export SERVER_URL=https://demo.dataverse.org
+  export API_KEY=xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx
+  export JSON={"url":"https://demo.dataverse.org/api/v1/datasets/:persistentId/?persistentId=doi:10.5072/FK2/J8SJZB","timeOut":5,"user":"alberteinstein"}
+
+  curl -H 'X-Dataverse-key:$API_KEY' -d $JSON $SERVER_URL/api/admin/requestSignedUrl
+
+Please see :ref:`dataverse.api.signature-secret` for the configuration option to add a shared secret, enabling extra
+security.

doc/sphinx-guides/source/installation/config.rst

@@ -580,8 +580,7 @@ Optionally, you may provide static credentials for each S3 storage using MicroPr
 - ``dataverse.files.<id>.access-key`` for this storage's "access key ID"
 - ``dataverse.files.<id>.secret-key`` for this storage's "secret access key"
 
-You may provide the values for these via any of the
-`supported config sources <https://docs.payara.fish/community/docs/documentation/microprofile/config/README.html>`_.
+You may provide the values for these via any `supported MicroProfile Config API source`_.
 
 **WARNING:**
 
@@ -1670,6 +1669,39 @@ This setting is useful in cases such as running your Dataverse installation behi
 	"HTTP_VIA",
 	"REMOTE_ADDR"
 
+
+.. _dataverse.api.signature-secret:
+
+dataverse.api.signature-secret
+++++++++++++++++++++++++++++++
+
+Context: Dataverse has the ability to create "Signed URLs" for it's API calls. Using a signed URLs is more secure than
+providing API tokens, which are long-lived and give the holder all of the permissions of the user. In contrast, signed URLs
+are time limited and only allow the action of the API call in the URL. See :ref:`api-exttools-auth` and
+:ref:`api-native-signed-url` for more details. 
+
+The key used to sign a URL is created from the API token of the creating user plus a signature-secret provided by an administrator.
+**Using a signature-secret is highly recommended.** This setting defaults to an empty string. Using a non-empty 
+signature-secret makes it impossible for someone who knows an API token from forging signed URLs and provides extra security by 
+making the overall signing key longer.
+
+Since the signature-secret is sensitive, you should treat it like a password. Here is an example how to set your shared secret 
+with the secure method "password alias":
+
+.. code-block:: shell
+
+  echo "AS_ADMIN_ALIASPASSWORD=change-me-super-secret" > /tmp/password.txt
+  asadmin create-password-alias --passwordfile /tmp/password.txt dataverse.api.signature-secret
+  rm /tmp/password.txt
+
+Can also be set via any `supported MicroProfile Config API source`_, e.g. the environment variable
+``DATAVERSE_API_SIGNATURE_SECRET``.
+
+**WARNING:** For security, do not use the sources "environment variable" or "system property" (JVM option) in a
+production context! Rely on password alias, secrets directory or cloud based sources instead!
+
+
+
 .. _:ApplicationServerSettings:
 
 Application Server Settings
@@ -3067,3 +3099,7 @@ The interval in seconds between Dataverse calls to Globus to check on upload pro
 +++++++++++++++++++++++++
 
 A true/false option to add a Globus transfer option to the file download menu which is not yet fully supported in the dataverse-globus app. See :ref:`globus-support` for details.
+
+
+
+.. _supported MicroProfile Config API source: https://docs.payara.fish/community/docs/Technical%20Documentation/MicroProfile/Config/Overview.html

src/main/java/edu/harvard/iq/dataverse/api/AbstractApiBean.java

@@ -46,9 +46,11 @@
 import edu.harvard.iq.dataverse.privateurl.PrivateUrlServiceBean;
 import edu.harvard.iq.dataverse.locality.StorageSiteServiceBean;
 import edu.harvard.iq.dataverse.search.savedsearch.SavedSearchServiceBean;
+import edu.harvard.iq.dataverse.settings.JvmSettings;
 import edu.harvard.iq.dataverse.settings.SettingsServiceBean;
 import edu.harvard.iq.dataverse.util.BundleUtil;
 import edu.harvard.iq.dataverse.util.SystemConfig;
+import edu.harvard.iq.dataverse.util.UrlSignerUtil;
 import edu.harvard.iq.dataverse.util.json.JsonParser;
 import edu.harvard.iq.dataverse.util.json.NullSafeJsonBuilder;
 import edu.harvard.iq.dataverse.validation.PasswordValidatorServiceBean;
@@ -362,7 +364,7 @@ protected AuthenticatedUser findUserByApiToken( String apiKey ) {
     protected User findUserOrDie() throws WrappedResponse {
         final String requestApiKey = getRequestApiKey();
         final String requestWFKey = getRequestWorkflowInvocationID();
-        if (requestApiKey == null && requestWFKey == null) {
+        if (requestApiKey == null && requestWFKey == null && getRequestParameter(UrlSignerUtil.SIGNED_URL_TOKEN)==null) {
             return GuestUser.get();
         }
         PrivateUrlUser privateUrlUser = privateUrlSvc.getPrivateUrlUserFromToken(requestApiKey);
@@ -419,10 +421,36 @@ private AuthenticatedUser findAuthenticatedUserOrDie( String key, String wfid )
             } else {
                 throw new WrappedResponse(badWFKey(wfid));
             }
+        } else if (getRequestParameter(UrlSignerUtil.SIGNED_URL_TOKEN) != null) {
+            AuthenticatedUser authUser = getAuthenticatedUserFromSignedUrl();
+            if (authUser != null) {
+                return authUser;
+            }
         }
         //Just send info about the apiKey - workflow users will learn about invocationId elsewhere
         throw new WrappedResponse(badApiKey(null));
     }
+
+    private AuthenticatedUser getAuthenticatedUserFromSignedUrl() {
+        AuthenticatedUser authUser = null;
+        // The signedUrl contains a param telling which user this is supposed to be for.
+        // We don't trust this. So we lookup that user, and get their API key, and use
+        // that as a secret in validating the signedURL. If the signature can't be
+        // validated with their key, the user (or their API key) has been changed and
+        // we reject the request.
+        // ToDo - add null checks/ verify that calling methods catch things.
+        String user = httpRequest.getParameter("user");
+        AuthenticatedUser targetUser = authSvc.getAuthenticatedUser(user);
+        String key = JvmSettings.API_SIGNING_SECRET.lookupOptional().orElse("")
+                + authSvc.findApiTokenByUser(targetUser).getTokenString();
+        String signedUrl = httpRequest.getRequestURL().toString() + "?" + httpRequest.getQueryString();
+        String method = httpRequest.getMethod();
+        boolean validated = UrlSignerUtil.isValidUrl(signedUrl, user, method, key);
+        if (validated) {
+            authUser = targetUser;
+        }
+        return authUser;
+    }
 
     protected Dataverse findDataverseOrDie( String dvIdtf ) throws WrappedResponse {
         Dataverse dv = findDataverse(dvIdtf);