Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Integrate daterange into datasets API #3371

Merged
merged 4 commits into from
Apr 10, 2023
Merged

Integrate daterange into datasets API #3371

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
520f9ec
Integrate daterange into datasets API
dbutenhof Apr 6, 2023
4e79e67
Allow both summary modes
dbutenhof Apr 6, 2023
fb5034a
A few more tweaks
dbutenhof Apr 7, 2023
510e264
Typo
dbutenhof Apr 7, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
193 changes: 105 additions & 88 deletions docs/API/V1/list.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -11,6 +11,11 @@ and arbitrary metadata filter expressions.
Large collections can be paginated for efficiency using the `limit` and `offset`
query parameters.

The `keysummary` and `daterange` query parameters (if `true`) select "summary"
modes where aggregate metadata is returned without a list of datasets. These two
may be used together, but cannot be used along with the normal collection list
mode as they aren't subject to pagination.

## Query parameters

`access` string \
Expand Down Expand Up @@ -65,6 +70,15 @@ case, a boolean, which is represented in JSON as `true` or `false`). Beware
especially when attempting to match a JSON document (such as
`dataset.metalog.pbench`).

`keysummary` boolean \
Instead of displaying a list of selected datasets and metadata, use the set of
specified filters to accumulate a nested report on the metadata key namespace
for the set of datasets. See [metadata](../metadata.md) for deails on the
Pbench Server metadata namespaces. Because the `global` and `user` namespaces
are completely dynamic, and the `dataset.metalog` sub-namespace varies greatly
across Pbench Agent benchmark scripts, this mode provides a mechanism for a
metadata visualizer to understand what's available for a set of datasets.

`limit` integer \
"Paginate" the selected datasets by returning at most `limit` datasets. This
can be used in conjunction with `offset` to progress through the full list in
Expand Down Expand Up @@ -102,15 +116,6 @@ If the timezone offset is omitted it will be assumed to be UTC (`+00:00`); if
the time is omitted it will be assumed as midnight (`00:00:00`) on the
specified date.

`keysummary` boolean \
Instead of displaying a list of selected datasets and metadata, use the set of
specified filters to accumulate a nested report on the metadata key namespace
for the set of datasets. See [metadata](../metadata.md) for deails on the
Pbench Server metadata namespaces. Because the `global` and `user` namespaces
are completely dynamic, and the `dataset.metalog` sub-namespace varies greatly
across Pbench Agent benchmark scripts, this mode provides a mechanism for a
metadata visualizer to understand what's available for a set of datasets.

## Request headers

`authorization: bearer` token [_optional_] \
Expand Down Expand Up @@ -155,7 +160,9 @@ a message, and optional JSON data provided by the system administrator.
### Dataset date range

The `application/json` response body is a JSON object describing the earliest
and most recent dataset upload time for the selected list of datasets.
and most recent dataset upload time for the selected list of datasets. If the
collection filters exclude all datasets (the result set is empty), the return
value will be empty, omitting both the `from` and `to` keywords.

```json
{
Expand Down Expand Up @@ -234,9 +241,9 @@ might return:

When the `keysummary` query parameter is `true` (e.g., either `?keysummary` or
`?keysummary=true`), instead of reporting a list of datasets and metadata for
each dataset, report a hierarchical representation of the aggregate metadata
namespace across all selected datasets. This returns much less data and is not
subject to pagination.
each dataset, the `application/json` response body contains a hierarchical
representation of the aggregate metadata namespace across all selected datasets.
This returns much less data and is not subject to pagination.

"Leaf" nodes in the metadata tree are represented by `null` values while any
key with children will be represented as a nested JSON object showing those
Expand All @@ -252,83 +259,93 @@ will return a JSON document with the keys `config`, `date`, `hostname_f`,

```json
{
"dataset": {
"access": null,
"id": null,
"metalog": {
"controller": {
"hostname": null,
"hostname-alias": null,
"hostname-all-fqdns": null,
"hostname-all-ip-addresses": null,
"hostname-domain": null,
"hostname-fqdn": null,
"hostname-ip-address": null,
"hostname-nis": null,
"hostname-short": null,
"ssh_opts": null
},
"iterations/1-default": {
"iteration_name": null,
"iteration_number": null,
"user_script": null
},
"pbench": {
"config": null,
"date": null,
"hostname_f": null,
"hostname_ip": null,
"hostname_s": null,
"iterations": null,
"name": null,
"rpm-version": null,
"script": null,
"tar-ball-creation-timestamp": null
},
"run": {
"controller": null,
"end_run": null,
"raw_size": null,
"start_run": null
"keys": {
"dataset": {
"access": null,
"id": null,
"metalog": {
"controller": {
"hostname": null,
"hostname-alias": null,
"hostname-all-fqdns": null,
"hostname-all-ip-addresses": null,
"hostname-domain": null,
"hostname-fqdn": null,
"hostname-ip-address": null,
"hostname-nis": null,
"hostname-short": null,
"ssh_opts": null
},
"iterations/1-default": {
"iteration_name": null,
"iteration_number": null,
"user_script": null
},
"pbench": {
"config": null,
"date": null,
"hostname_f": null,
"hostname_ip": null,
"hostname_s": null,
"iterations": null,
"name": null,
"rpm-version": null,
"script": null,
"tar-ball-creation-timestamp": null
},
"run": {
"controller": null,
"end_run": null,
"raw_size": null,
"start_run": null
},
"tools": {
"group": null,
"hosts": null,
"trigger": null
},
"tools/dbutenho.bos.csb": {
"hostname-alias": null,
"hostname-all-fqdns": null,
"hostname-all-ip-addresses": null,
"hostname-domain": null,
"hostname-fqdn": null,
"hostname-ip-address": null,
"hostname-nis": null,
"hostname-short": null,
"label": null,
"rpm-version": null,
"tools": null,
"vmstat": null
},
"tools/dbutenho.bos.csb/vmstat": {
"install_check_output": null,
"install_check_status_code": null,
"options": null
}
},
"tools": {
"group": null,
"hosts": null,
"trigger": null
},
"tools/dbutenho.bos.csb": {
"hostname-alias": null,
"hostname-all-fqdns": null,
"hostname-all-ip-addresses": null,
"hostname-domain": null,
"hostname-fqdn": null,
"hostname-ip-address": null,
"hostname-nis": null,
"hostname-short": null,
"label": null,
"rpm-version": null,
"tools": null,
"vmstat": null
},
"tools/dbutenho.bos.csb/vmstat": {
"install_check_output": null,
"install_check_status_code": null,
"options": null
}
"name": null,
"owner_id": null,
"resource_id": null,
"uploaded": null
},
"name": null,
"owner_id": null,
"resource_id": null,
"uploaded": null
},
"server": {
"deletion": null,
"index-map": {
"container-pbench.v6.run-data.2023-03": null,
"container-pbench.v6.run-toc.2023-03": null
},
"origin": null,
"tarball-path": null
"server": {
"deletion": null,
"index-map": {
"container-pbench.v6.run-data.2023-03": null,
"container-pbench.v6.run-toc.2023-03": null
},
"origin": null,
"tarball-path": null
}
}
}
```

### Combining key namespace summary and date range

When both the `keysummary` and `daterange` query parameters are `true`, the
`application/json` response body contains the `from`, `to`, and `keys` key
values. If the selected collection filters produce no results, as with
`daterange` alone, the `from` and `to` keys will be omitted and the value of
`keys` will be an empty object.
webbnh marked this conversation as resolved.
Show resolved Hide resolved
32 changes: 22 additions & 10 deletions lib/pbench/server/api/resources/datasets_list.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -347,18 +347,16 @@ def keyspace(self, query: Query) -> JSONOBJECT:
self.accumulate(aggregate["dataset"], m.key, m.value)
else:
self.accumulate(aggregate, m.key, m.value)
return aggregate
return {"keys": aggregate}

def daterange(self, request: Request, json: JSONOBJECT, query: Query) -> JSONOBJECT:
def daterange(self, query: Query) -> JSONOBJECT:
"""Return only the date range of the selected datasets.

Replace the selected "entities" (normally Dataset columns) with the
SQL min and max functions on the dataset upload timestamp so that the
generated SQL query will return a tuple of those two values.

Args:
request: The HTTP Request object
json: The JSON query parameters
query: The basic filtered SQLAlchemy query object

Returns:
Expand All @@ -368,7 +366,10 @@ def daterange(self, request: Request, json: JSONOBJECT, query: Query) -> JSONOBJ
func.min(Dataset.uploaded), func.max(Dataset.uploaded)
).first()

return {"from": results[0].isoformat(), "to": results[1].isoformat()}
if results and results[0] and results[1]:
return {"from": results[0].isoformat(), "to": results[1].isoformat()}
else:
return {}

def datasets(self, request: Request, json: JSONOBJECT, query: Query) -> JSONOBJECT:
"""Gather and paginate the selected datasets
Expand Down Expand Up @@ -517,9 +518,20 @@ def _get(
else:
owner = json.get("owner")
query = self._build_sql_query(owner, json.get("access"), query)
result = {}
done = False

# We can do "keysummary" and "daterange", but, as it makes no real
# sense to paginate either, we don't support them in combination with
# a normal list query. So we will perform either/or keysummary and
# daterange, and acquire a normal list of datasets only if neither was
# specified.
if json.get("keysummary"):
return jsonify(self.keyspace(query))
elif json.get("daterange"):
return jsonify(self.daterange(request, json, query))
else:
return jsonify(self.datasets(request, json, query))
result.update(self.keyspace(query))
done = True
if json.get("daterange"):
result.update(self.daterange(query))
done = True
if not done:
result = self.datasets(request, json, query)
return jsonify(result)
Loading