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.

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

Add delete all for tenants in Vespa #3970

Merged
merged 1 commit into from
Feb 13, 2025
Merged
Changes from all commits
Commits
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
178 changes: 170 additions & 8 deletions backend/scripts/debugging/onyx_vespa.py
Original file line number Diff line number Diff line change
Expand Up @@ -256,16 +256,28 @@ def get_documents_for_tenant_connector(


def search_for_document(
index_name: str, document_id: str, max_hits: int | None = 10
index_name: str,
document_id: str | None = None,
tenant_id: str | None = None,
max_hits: int | None = 10,
) -> List[Dict[str, Any]]:
yql_query = (
f'select * from sources {index_name} where document_id contains "{document_id}"'
)
yql_query = f"select * from sources {index_name}"

conditions = []
if document_id is not None:
conditions.append(f'document_id contains "{document_id}"')

if tenant_id is not None:
conditions.append(f'tenant_id contains "{tenant_id}"')

if conditions:
yql_query += " where " + " and ".join(conditions)

params: dict[str, Any] = {"yql": yql_query}
if max_hits is not None:
params["hits"] = max_hits
with get_vespa_http_client() as client:
response = client.get(f"{SEARCH_ENDPOINT}/search/", params=params)
response = client.get(f"{SEARCH_ENDPOINT}search/", params=params)
response.raise_for_status()
result = response.json()
documents = result.get("root", {}).get("children", [])
Expand Down Expand Up @@ -582,8 +594,15 @@ def update_document(
) -> None:
update_document(self.tenant_id, connector_id, doc_id, fields)

def search_for_document(self, document_id: str) -> List[Dict[str, Any]]:
return search_for_document(self.index_name, document_id)
def delete_documents_for_tenant(self, count: int | None = None) -> None:
if not self.tenant_id:
raise Exception("Tenant ID is not set")
delete_documents_for_tenant(self.index_name, self.tenant_id, count=count)

def search_for_document(
self, document_id: str | None = None, tenant_id: str | None = None
) -> List[Dict[str, Any]]:
return search_for_document(self.index_name, document_id, tenant_id)

def delete_document(self, connector_id: int, doc_id: str) -> None:
# Delete a document.
Expand All @@ -600,6 +619,147 @@ def acls(self, cc_pair_id: int, n: int | None = 10) -> None:
get_document_acls(self.tenant_id, cc_pair_id, n)


def delete_where(
index_name: str,
selection: str,
cluster: str = "default",
bucket_space: str | None = None,
continuation: str | None = None,
time_chunk: str | None = None,
timeout: str | None = None,
tracelevel: int | None = None,
) -> None:
"""
Removes visited documents in `cluster` where the given selection
is true, using Vespa's 'delete where' endpoint.

:param index_name: Typically <namespace>/<document-type> from your schema
:param selection: The selection string, e.g., "true" or "foo contains 'bar'"
:param cluster: The name of the cluster where documents reside
:param bucket_space: e.g. 'global' or 'default'
:param continuation: For chunked visits
:param time_chunk: If you want to chunk the visit by time
:param timeout: e.g. '10s'
:param tracelevel: Increase for verbose logs
"""
# Using index_name of form <namespace>/<document-type>, e.g. "nomic_ai_nomic_embed_text_v1"
# This route ends with "/docid/" since the actual ID is not specified — we rely on "selection".
path = f"/document/v1/{index_name}/docid/"

params = {
"cluster": cluster,
"selection": selection,
}

# Optional parameters
if bucket_space is not None:
params["bucketSpace"] = bucket_space
if continuation is not None:
params["continuation"] = continuation
if time_chunk is not None:
params["timeChunk"] = time_chunk
if timeout is not None:
params["timeout"] = timeout
if tracelevel is not None:
params["tracelevel"] = tracelevel # type: ignore

with get_vespa_http_client() as client:
url = f"{VESPA_APPLICATION_ENDPOINT}{path}"
logger.info(f"Performing 'delete where' on {url} with selection={selection}...")
response = client.delete(url, params=params)
# (Optionally, you can keep fetching `continuation` from the JSON response
# if you have more documents to delete in chunks.)
response.raise_for_status() # will raise HTTPError if not 2xx
logger.info(f"Delete where completed with status: {response.status_code}")
print(f"Delete where completed with status: {response.status_code}")


def delete_documents_for_tenant(
index_name: str,
tenant_id: str,
route: str | None = None,
condition: str | None = None,
timeout: str | None = None,
tracelevel: int | None = None,
count: int | None = None,
) -> None:
"""
For the given tenant_id and index_name (often in the form <namespace>/<document-type>),
find documents via search_for_document, then delete them one at a time using Vespa's
/document/v1/<namespace>/<document-type>/docid/<document-id> endpoint.

:param index_name: Typically <namespace>/<document-type> from your schema
:param tenant_id: The tenant to match in your Vespa search
:param route: Optional route parameter for delete
:param condition: Optional conditional remove
:param timeout: e.g. '10s'
:param tracelevel: Increase for verbose logs
"""
deleted_count = 0
while True:
# Search for documents with the given tenant_id
docs = search_for_document(
index_name=index_name,
document_id=None,
tenant_id=tenant_id,
max_hits=100, # Fetch in batches of 100
)

if not docs:
logger.info("No more documents found to delete.")
break

with get_vespa_http_client() as client:
for doc in docs:
if count is not None and deleted_count >= count:
logger.info(f"Reached maximum delete limit of {count} documents.")
return

fields = doc.get("fields", {})
doc_id_value = fields.get("document_id") or fields.get("documentid")
tenant_id = fields.get("tenant_id")
if tenant_id != tenant_id:
raise Exception("Tenant ID mismatch")

if not doc_id_value:
logger.warning(
"Skipping a document that has no document_id in 'fields'."
)
continue

url = f"{DOCUMENT_ID_ENDPOINT.format(index_name=index_name)}/{doc_id_value}"

params = {}
if condition:
params["condition"] = condition
if route:
params["route"] = route
if timeout:
params["timeout"] = timeout
if tracelevel is not None:
params["tracelevel"] = str(tracelevel)

response = client.delete(url, params=params)
if response.status_code == 200:
logger.info(f"Successfully deleted doc_id={doc_id_value}")
deleted_count += 1
else:
logger.error(
f"Failed to delete doc_id={doc_id_value}, "
f"status={response.status_code}, response={response.text}"
)
print(
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Any reason to print if we are already using logger.error?

f"Could not delete doc_id={doc_id_value}. "
f"Status={response.status_code}, response={response.text}"
)
raise Exception(
f"Could not delete doc_id={doc_id_value}. "
f"Status={response.status_code}, response={response.text}"
)

logger.info(f"Deleted {deleted_count} documents in total.")


def main() -> None:
parser = argparse.ArgumentParser(description="Vespa debugging tool")
parser.add_argument(
Expand Down Expand Up @@ -630,7 +790,9 @@ def main() -> None:
args = parser.parse_args()
vespa_debug = VespaDebugging(args.tenant_id)

if args.action == "config":
if args.action == "delete-all-documents":
vespa_debug.delete_documents_for_tenant(args.count)
pablonyx marked this conversation as resolved.
Show resolved Hide resolved
elif args.action == "config":
vespa_debug.print_config()
elif args.action == "connect":
vespa_debug.check_connectivity()
Expand Down
Loading