Skip to content
For the complete documentation index, see llms.txt. This page is also available as Markdown.

Inspect stats

Read row counts, index state, and partitions from a collection's stats, alongside its name, key column, and external IDs.

stats() returns one dictionary describing a collection: the runtime statistics — row counts, index state, partitions — plus the collection's name, its partition-key column key_name, and its external_ids. Use it to read the size and shape of a collection without reading any records.

stats() reports live state only; it does not read payloads. It works whether the collection uses exact or IVF search.

Read the stats

from telys import Telys

telys = Telys("./telys-data")
col = telys.open_collection("notes")   # no embedder needed to read stats

s = col.stats()

print("name       :", s["name"])
print("key column :", s["key_name"])
print("external   :", s["external_ids"])
print("full stats :", s)

What the dictionary contains

Key
Always present
What it tells you

name

yes

The collection name.

key_name

yes

The physical partition-key column (from partition_by).

external_ids

yes

The live external IDs — mirrors len(col).

runtime stats

runtime-provided

Row counts, index state (exact / IVF), partition breakdown, and probe settings.

key_name is the single physical partition key. A list or tuple passed to partition_by uses only its first element — Telys builds no composite keys, so compose one string with scope_key(...) if you need several parts. See Partitions & scope keys.

Expected result

For a saved collection of 1,240 rows partitioned by scope, still on exact search:

external_ids always agrees with len(col). If index reports exact on a large collection, that is your cue to build an IVF index or run a tuning pass.

stats() is also available over a RemoteTelys connection — RemoteCollection includes it, so you can inspect a served collection the same way.

Next steps

Last updated

Was this helpful?