Datasets:

csoai
/

eu-ecosystem-measurements

Tags:

License:

Dataset card Files Files and versions

xet

Community

Dataset Viewer

Duplicate

Split (1)

train

The dataset viewer is not available for this split.

Cannot load the dataset split (in streaming mode) to extract the first rows.

Error code:   StreamingRowsError
Exception:    CastError
Message:      Couldn't cast
as_of: timestamp[s]
asset_endpoint_state_counts: struct<ENDPOINT_RESPONDS: int64>
  child 0, ENDPOINT_RESPONDS: int64
asset_gaiax_serviceSD_http_state_counts: struct<ENDPOINT_RESPONDS: int64, ENDPOINT_RESPONDS_5XX: int64, ENDPOINT_UNREACHABLE: int64, NOT_DECL (... 32 chars omitted)
  child 0, ENDPOINT_RESPONDS: int64
  child 1, ENDPOINT_RESPONDS_5XX: int64
  child 2, ENDPOINT_UNREACHABLE: int64
  child 3, NOT_DECLARED: int64
  child 4, UNCHECKABLE: int64
asset_types_declared: struct<SaaS: int64, access: int64, algorithm: int64, compute: int64, dataset: int64, service: int64>
  child 0, SaaS: int64
  child 1, access: int64
  child 2, algorithm: int64
  child 3, compute: int64
  child 4, dataset: int64
  child 5, service: int64
assets_read: int64
catalogue: struct<docs_returned: int64, duplicate_ids: int64, note_on_duplicates: string, pages: list<item: str (... 224 chars omitted)
  child 0, docs_returned: int64
  child 1, duplicate_ids: int64
  child 2, note_on_duplicates: string
  child 3, pages: list<item: struct<bytes: int64, from: int64, http_status: int64, sha256: string, size: int64>>
      child 0, item: struct<bytes: int64, from: int64, http_status: int64, sha256: string, size: int64>
          child 0, bytes: int64
          child 1, from: int64
          child 2, http_status: int64
          child 3, sha256: string
          child 4, size: int64
  child 4, query: struct<exists: struct<field: string>>
      child 0, exists: struct<field: string>
          child 0, fiel
...
64, lines: int64, sha256: string>
      child 0, bytes: int64
      child 1, lines: int64
      child 2, sha256: string
  child 3, records.jsonl: struct<bytes: int64, lines: int64, sha256: string>
      child 0, bytes: int64
      child 1, lines: int64
      child 2, sha256: string
withheld_files_note: string
what_this_is: string
proofs: list<item: struct<original: string, original_oid: string, original_sha256: string, upgraded: string, (... 246 chars omitted)
  child 0, item: struct<original: string, original_oid: string, original_sha256: string, upgraded: string, upgraded_s (... 234 chars omitted)
      child 0, original: string
      child 1, original_oid: string
      child 2, original_sha256: string
      child 3, upgraded: string
      child 4, upgraded_sha256: string
      child 5, file_digest: string
      child 6, target_file: string
      child 7, target_digest_matches_proof: bool
      child 8, bitcoin: list<item: struct<height: int64, block_hash: string, block_time: int64, merkle_root: string, header_ (... 32 chars omitted)
          child 0, item: struct<height: int64, block_hash: string, block_time: int64, merkle_root: string, header_source: str (... 20 chars omitted)
              child 0, height: int64
              child 1, block_hash: string
              child 2, block_time: int64
              child 3, merkle_root: string
              child 4, header_source: string
              child 5, result: string
how_to_verify: string
date: timestamp[s]
repo: string
to
{'schema': Value('string'), 'date': Value('timestamp[s]'), 'repo': Value('string'), 'what_this_is': Value('string'), 'how_to_verify': Value('string'), 'proofs': List({'original': Value('string'), 'original_oid': Value('string'), 'original_sha256': Value('string'), 'upgraded': Value('string'), 'upgraded_sha256': Value('string'), 'file_digest': Value('string'), 'target_file': Value('string'), 'target_digest_matches_proof': Value('bool'), 'bitcoin': List({'height': Value('int64'), 'block_hash': Value('string'), 'block_time': Value('int64'), 'merkle_root': Value('string'), 'header_source': Value('string'), 'result': Value('string')})})}
because column names don't match
Traceback:    Traceback (most recent call last):
                File "/src/services/worker/src/worker/utils.py", line 147, in get_rows_or_raise
                  return get_rows(
                      dataset=dataset,
                  ...<4 lines>...
                      column_names=column_names,
                  )
                File "/src/libs/libcommon/src/libcommon/utils.py", line 272, in decorator
                  return func(*args, **kwargs)
                File "/src/services/worker/src/worker/utils.py", line 127, in get_rows
                  rows_plus_one = list(itertools.islice(safe_iter(ds, dataset=dataset), rows_max_number + 1))
                File "/src/services/worker/src/worker/utils.py", line 483, in safe_iter
                  yield from ds.decode(False) if ds.features else ds
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2951, in __iter__
                  for key, example in ex_iterable:
                                      ^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2461, in __iter__
                  for key, pa_table in self._iter_arrow():
                                       ~~~~~~~~~~~~~~~~^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2486, in _iter_arrow
                  for key, pa_table in self.ex_iterable._iter_arrow():
                                       ~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 547, in _iter_arrow
                  for key, pa_table in iterator:
                                       ^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 430, in _iter_arrow
                  for key, pa_table in self.generate_tables_fn(**gen_kwags):
                                       ~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 343, in _generate_tables
                  self._cast_table(pa_table, json_field_paths=json_field_paths),
                  ~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 132, in _cast_table
                  pa_table = table_cast(pa_table, self.info.features.arrow_schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2378, in table_cast
                  return cast_table_to_schema(table, schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2306, in cast_table_to_schema
                  raise CastError(
                  ...<3 lines>...
                  )
              datasets.table.CastError: Couldn't cast
              as_of: timestamp[s]
              asset_endpoint_state_counts: struct<ENDPOINT_RESPONDS: int64>
                child 0, ENDPOINT_RESPONDS: int64
              asset_gaiax_serviceSD_http_state_counts: struct<ENDPOINT_RESPONDS: int64, ENDPOINT_RESPONDS_5XX: int64, ENDPOINT_UNREACHABLE: int64, NOT_DECL (... 32 chars omitted)
                child 0, ENDPOINT_RESPONDS: int64
                child 1, ENDPOINT_RESPONDS_5XX: int64
                child 2, ENDPOINT_UNREACHABLE: int64
                child 3, NOT_DECLARED: int64
                child 4, UNCHECKABLE: int64
              asset_types_declared: struct<SaaS: int64, access: int64, algorithm: int64, compute: int64, dataset: int64, service: int64>
                child 0, SaaS: int64
                child 1, access: int64
                child 2, algorithm: int64
                child 3, compute: int64
                child 4, dataset: int64
                child 5, service: int64
              assets_read: int64
              catalogue: struct<docs_returned: int64, duplicate_ids: int64, note_on_duplicates: string, pages: list<item: str (... 224 chars omitted)
                child 0, docs_returned: int64
                child 1, duplicate_ids: int64
                child 2, note_on_duplicates: string
                child 3, pages: list<item: struct<bytes: int64, from: int64, http_status: int64, sha256: string, size: int64>>
                    child 0, item: struct<bytes: int64, from: int64, http_status: int64, sha256: string, size: int64>
                        child 0, bytes: int64
                        child 1, from: int64
                        child 2, http_status: int64
                        child 3, sha256: string
                        child 4, size: int64
                child 4, query: struct<exists: struct<field: string>>
                    child 0, exists: struct<field: string>
                        child 0, fiel
              ...
              64, lines: int64, sha256: string>
                    child 0, bytes: int64
                    child 1, lines: int64
                    child 2, sha256: string
                child 3, records.jsonl: struct<bytes: int64, lines: int64, sha256: string>
                    child 0, bytes: int64
                    child 1, lines: int64
                    child 2, sha256: string
              withheld_files_note: string
              what_this_is: string
              proofs: list<item: struct<original: string, original_oid: string, original_sha256: string, upgraded: string, (... 246 chars omitted)
                child 0, item: struct<original: string, original_oid: string, original_sha256: string, upgraded: string, upgraded_s (... 234 chars omitted)
                    child 0, original: string
                    child 1, original_oid: string
                    child 2, original_sha256: string
                    child 3, upgraded: string
                    child 4, upgraded_sha256: string
                    child 5, file_digest: string
                    child 6, target_file: string
                    child 7, target_digest_matches_proof: bool
                    child 8, bitcoin: list<item: struct<height: int64, block_hash: string, block_time: int64, merkle_root: string, header_ (... 32 chars omitted)
                        child 0, item: struct<height: int64, block_hash: string, block_time: int64, merkle_root: string, header_source: str (... 20 chars omitted)
                            child 0, height: int64
                            child 1, block_hash: string
                            child 2, block_time: int64
                            child 3, merkle_root: string
                            child 4, header_source: string
                            child 5, result: string
              how_to_verify: string
              date: timestamp[s]
              repo: string
              to
              {'schema': Value('string'), 'date': Value('timestamp[s]'), 'repo': Value('string'), 'what_this_is': Value('string'), 'how_to_verify': Value('string'), 'proofs': List({'original': Value('string'), 'original_oid': Value('string'), 'original_sha256': Value('string'), 'upgraded': Value('string'), 'upgraded_sha256': Value('string'), 'file_digest': Value('string'), 'target_file': Value('string'), 'target_digest_matches_proof': Value('bool'), 'bitcoin': List({'height': Value('int64'), 'block_hash': Value('string'), 'block_time': Value('int64'), 'merkle_root': Value('string'), 'header_source': Value('string'), 'result': Value('string')})})}
              because column names don't match

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

EU ecosystem measurements

Start here — free verify: https://councilof.ai/gspc-verify/ (or the free MCP tool verify_card at https://councilof.ai/mcp) · Full catalog: https://councilof.ai/catalog.json · Methodology: https://councilof.ai/methodology/ · Measurement, not certification.

This dataset is published by the Council of AI (CSOAI), an independent measurement layer. We measure. We never certify. No organisation named here is a client, partner or member of anything of ours, and none of them has reviewed or approved this work.

Release 1: Pontus-X public catalogue, declared vs observed (2026-09-27), aggregate only

Files are in pontus-x/2026-09-27/:

File	What it is
`summary.json`	The aggregate counts (redacted, see below).
`summary.signed.json`	The Ed25519 signature over `summary.json`, from `did:web:csoai.org#board-attestation-1`.
`summary.json.ots`	The OpenTimestamps proof for the `summary.json` bytes.
`summary.ots.json`	The state of that proof.

Method

Read-only. The source is the public Aquarius metadata catalogue at https://aquarius.pontus-x.eu/api/aquarius/assets/query, for chains 32456 and 32457. We queried every document that has metadata. The query spans 12 Elasticsearch indices: 9,591 documents, which are 3,403 unique DIDs.
Endpoints and metadata only. For each distinct URL a document declares, we made one probe and recorded:
- DNS resolution;
- TLS certificate validity;
- the HTTP status of a HEAD request, or of a GET whose body was not read.
No data consumed. Nothing was ordered, purchased, downloaded or consumed. The file URLs are encrypted and served only after an order, so file availability was not examined.
Politeness. robots.txt was honoured, at most 1 request started per host per second, UA CSOAI-measure/0.1 (+https://councilof.ai).
Timing. The run went from 2026-09-27T05:51:44Z to 05:53:06Z. Every URL was probed once, so there is no uptime measurement.

Counts (from `summary.json`)

Assets and declared service endpoints

Measure	Count
Unique assets (DIDs) read	3,403
Assets by chain	1,597 on chain 32456; 1,806 on chain 32457
Asset types declared	2,913 dataset, 483 algorithm, 4 compute, 1 SaaS, 1 access, 1 service
Distinct declared service endpoints	20
Service endpoints that responded	20 of 20
Assets whose declared service endpoint responded	3,403 of 3,403
Distinct service endpoints with a TLS finding	2: one expired certificate, one "unable to get local issuer certificate"
Per-asset declared services, by TLS state	8 TLS_INVALID; 3,398 TLS_VALID (3,406 services across 3,403 assets)

Declared Gaia-X serviceSD URLs

There are 418 distinct serviceSD URLs:

97 respond;
6 answer with a 5xx status;
314 are unreachable (in 310 the DNS does not resolve, and 4 resolve but give no HTTP response);
1 is UNCHECKABLE because it is not an http(s) URL.

By asset, 2,882 declare no serviceSD URL.

What this shows

It shows whether each endpoint that a public catalogue entry declares resolved, answered and presented a valid TLS certificate at probe time.

What this does not show

Data delivery or file availability.
Whether any DID, credential or Gaia-X self-description verifies. The content of the serviceSD documents was not checked, so it is UNMEASURED.
Uptime.
Anything about the quality or conduct of any publisher, operator or participant, or of the ecosystem.

This is not a score, a ranking or a compliance determination, and no legal provision is bound to it.

Why this release is aggregate-only (redaction)

The signed staging summary named two service-endpoint hostnames (the two TLS findings) and one organisation in its disclaimer. This release names no operator. So:

this summary.json is a redacted aggregate: the hostnames are replaced by a count and the two reasons, and the organisation name is removed;
it was re-signed and re-timestamped as its own artifact;
every count is unchanged.

The unredacted summary is committed here by hash only (redaction.source_summary_sha256), and so are the per-asset and per-URL files (withheld_files). None of them is published in this release, because they identify operators, DIDs and endpoint URLs.

How to verify

import json, hashlib, base64, urllib.request
from cryptography.hazmat.primitives.asymmetric import ed25519
s = json.load(open("summary.signed.json")); raw = open("summary.json", "rb").read()
c = json.dumps(s["payload"], sort_keys=True, separators=(",", ":"), ensure_ascii=False).encode()
assert hashlib.sha256(raw).hexdigest() == s["payload"]["artifact"]["sha256"]   # file is the signed one
assert hashlib.sha256(c).hexdigest() == s["signature"]["payload_sha256"]       # payload preimage
req = urllib.request.Request("https://csoai.org/.well-known/did.json", headers={"user-agent": "Mozilla/5.0"})
did = json.load(urllib.request.urlopen(req))
x = [m for m in did["verificationMethod"] if m["id"].endswith("#board-attestation-1")][0]["publicKeyJwk"]["x"]
ed25519.Ed25519PublicKey.from_public_bytes(base64.urlsafe_b64decode(x + "==")).verify(
    bytes.fromhex(s["signature"]["sig_ed25519"]), c)
print("VERIFIES")

For the timestamp, run pip install opentimestamps-client, then ots upgrade summary.json.ots and then ots verify summary.json.ots.

At publication the proof is a pending calendar commitment, not a Bitcoin attestation. It becomes one only after an upgrade returns a Bitcoin block-header attestation.

The signature proves only that these bytes were signed by the board key at signature.signed_at. It proves nothing inside the record beyond what the record's own instruments measured.

Objections and corrections

If you operate anything in this catalogue and disagree with a count, or want context added, please use https://councilof.ai/census/ to tell us. Corrections are published as new dated files. They never silently overwrite.

Objections, contact and corrections

To object to a row, ask for a re-check, request a correction or ask for a record to be withdrawn, email nicholas@csoai.org or use the appeals and dispute route at https://councilof.ai/dispute/. Corrections are listed in the corrections ledger at https://councilof.ai/api/corrections, with what changed and when. CSOAI Ltd (company no. 16939677, England and Wales) is the accountable publisher.

How to cite

CSOAI Ltd (Council of AI). EU ecosystem measurements (aggregate, signed). 2026. Hugging Face dataset csoai/eu-ecosystem-measurements. https://huggingface.co/datasets/csoai/eu-ecosystem-measurements

@misc{csoai_eu_ecosystem_measurements,
  title        = {EU ecosystem measurements (aggregate, signed)},
  author       = {{CSOAI Ltd}},
  year         = {2026},
  howpublished = {Hugging Face dataset, https://huggingface.co/datasets/csoai/eu-ecosystem-measurements},
  note         = {Corrections: https://councilof.ai/api/corrections}
}

Licence: CC-BY-4.0. Attribute Council of AI, CSOAI Ltd (16939677), https://councilof.ai.

Corrections and verification

Corrections ledger (signed): https://councilof.ai/api/corrections. Corrections to CSOAI's published records are logged there with what changed and when.
Verify a signed record yourself, free and without an account: https://councilof.ai/gspc-verify/ (step by step: https://councilof.ai/signed/HOW-TO-VERIFY.md).
Conformance kit for signed-receipts/v1, with test vectors for implementers: https://councilof.ai/spec/signed-receipts/v1/conformance/

Downloads last month: 158

Total file size:

25.5 kB