trodo-python 2.25.0__py3-none-any.whl → 2.26.1__py3-none-any.whl

trodo/__init__.py

@@ -41,7 +41,7 @@ Downstream microservice (join the caller's run instead of making a new one):
 
 from __future__ import annotations
 
-__version__ = "2.25.0"
+__version__ = "2.26.1"
 
 from typing import Any, Callable, Dict, List, Optional, Union
 
@@ -122,6 +122,7 @@ __all__ = [
     "PromptSummary",
     # Datasets & experiments
     "append_dataset",
+    "export_dataset",
     "ingest_experiment",
     "create_experiment",
     "run_experiment",
@@ -435,6 +436,30 @@ def append_dataset(
     return _get_client().datasets.append(ref, items, create=create)
 
 
+def export_dataset(
+    ref: str,
+    token: str,
+    labels: str = "all",
+    include_unlabelled: bool = False,
+    version: Optional[int] = None,
+    since: Optional[str] = None,
+) -> Dict[str, Any]:
+    """Read a dataset's labelled export with an export token (Settings → Datasets).
+
+    Rows carry ``input``, ``output``, ``expected``, ``evaluations``,
+    ``label_source`` / ``label_status`` / ``label_weight`` and ``source``::
+
+        out = trodo.export_dataset("support-failures", token=os.environ["TRODO_EXPORT_TOKEN"], labels="confirmed")
+        for row in out["rows"]:
+            ...
+
+    Returns ``{"rows": [...], "version": int | None}``.
+    """
+    return _get_client().datasets.export(
+        ref, token, labels=labels, include_unlabelled=include_unlabelled, version=version, since=since,
+    )
+
+
 def ingest_experiment(
     dataset: str,
     outputs: List[Dict[str, Any]],

trodo/_version.py

@@ -1 +1 @@
-SDK_VERSION = "2.25.0"
+SDK_VERSION = "2.26.0"

trodo/api/http_client.py

@@ -180,6 +180,45 @@ class HttpClient:
             f"/api/sdk/datasets/{quote(str(ref), safe='')}/items", body
         )
 
+    def export_dataset(self, ref: str, params: Dict[str, Any], token: str) -> Dict[str, Any]:
+        """Read a dataset's labelled export (GET /datasets/:ref/export) as JSONL
+        text, authenticated with an EXPORT TOKEN (``trodo_dx_…``), never the
+        site id. Returns ``{"text", "status", "rows", "version"}`` or
+        ``{"__error": True, "status", ...}``. No retries on 4xx."""
+        from urllib.parse import quote
+        url = f"{self.api_base}/api/sdk/datasets/{quote(str(ref), safe='')}/export"
+        clean = {k: v for k, v in (params or {}).items() if v is not None and v != ""}
+        self._log(f"GET {url}")
+        try:
+            resp = requests.get(
+                url,
+                params=clean,
+                headers={"Authorization": f"Bearer {token}", "User-Agent": f"trodo-python/{SDK_VERSION}"},
+                timeout=max(self.timeout, 60),
+            )
+        except Exception as exc:  # network
+            self._log(f"Error: {exc}")
+            if self.on_error:
+                self.on_error(exc)
+            return {"__error": True, "message": str(exc)}
+        if resp.status_code >= 400:
+            base: Dict[str, Any] = {"__error": True, "status": resp.status_code}
+            try:
+                data = resp.json()
+                if isinstance(data, dict):
+                    base.update(data)
+            except Exception:
+                base["message"] = resp.text[:300]
+            return base
+        rows = resp.headers.get("x-trodo-export-rows")
+        version = resp.headers.get("x-trodo-dataset-version")
+        return {
+            "text": resp.text,
+            "status": resp.status_code,
+            "rows": int(rows) if rows and rows.isdigit() else None,
+            "version": int(version) if version and version.isdigit() else None,
+        }
+
     def ingest_experiment(self, body: Dict[str, Any]) -> ApiResult:
         return self._request("/api/sdk/experiments/ingest", body)
 

trodo/managers/dataset_manager.py

@@ -12,20 +12,33 @@ Mirrors ``sdks/trodo-node-sdk/src/managers/DatasetManager.ts`` in shape.
 
 from __future__ import annotations
 
-from typing import Any, Dict, List, Union
+import json
+from typing import Any, Dict, List, Optional, Union
 
 __all__ = ["DatasetManager"]
 
-# The optional per-item fields, in wire order. ``input`` is always required.
-_ITEM_OPTIONAL = ("expected_output", "metadata", "source_trace_ref")
+# The optional per-item fields, in wire order. ``input`` is required on a row.
+_ITEM_OPTIONAL = ("output", "expected_output", "metadata", "source_trace_ref", "tags")
+# A reference to production instead of content: the item becomes that unit's
+# row (Input, Output and Evaluations filled by Trodo). ``eval_run_id`` names a
+# verdict: Output is what the eval said, Expected its correction.
+_UNIT_REF_KEYS = ("span_id", "run_id", "conversation_id", "eval_run_id")
+_UNIT_REF_OPTIONAL = ("tags", "metadata")
 
 
 def _clean_item(item: Dict[str, Any]) -> Dict[str, Any]:
     if not isinstance(item, dict):
         raise ValueError("trodo: each dataset item must be a dict")
+    ref = [k for k in _UNIT_REF_KEYS if item.get(k)]
+    if ref and ("input" not in item or item["input"] is None):
+        out: Dict[str, Any] = {ref[0]: str(item[ref[0]])}
+        for key in _UNIT_REF_OPTIONAL:
+            if item.get(key) is not None:
+                out[key] = item[key]
+        return out
     if "input" not in item or item["input"] is None:
-        raise ValueError("trodo: each dataset item requires an 'input' dict")
-    out: Dict[str, Any] = {"input": item["input"]}
+        raise ValueError("trodo: each dataset item requires an 'input' dict, or a span_id / run_id / conversation_id / eval_run_id")
+    out = {"input": item["input"]}
     for key in _ITEM_OPTIONAL:
         if item.get(key) is not None:
             out[key] = item[key]
@@ -72,3 +85,50 @@ class DatasetManager:
             "create": bool(create),
         }
         return self._http.append_dataset_items(ref, body)
+
+    def export(
+        self,
+        ref: str,
+        token: str,
+        labels: str = "all",
+        include_unlabelled: bool = False,
+        version: Optional[int] = None,
+        since: Optional[str] = None,
+    ) -> Dict[str, Any]:
+        """Read the dataset *ref*'s labelled export.
+
+        Each row is the dataset's training shape: ``input``, ``output``,
+        ``expected``, ``evaluations`` (the Evals v3 verdicts snapshotted onto
+        the row), ``label_source`` / ``label_status`` / ``label_weight``,
+        ``source``. *token* is an export token from Settings → Datasets
+        (``trodo_dx_…``); the site id cannot read exports. *labels* is one of
+        ``all, confirmed, explicit, derived, implicit, proposed, unlabelled,
+        failed_only, passed_only``; *version* pins a snapshot.
+
+        Returns ``{"rows": [...], "version": int | None}``.
+
+        :raises ValueError: if *ref* or *token* is empty.
+        :raises RuntimeError: on a non-2xx response (the message carries the status and body).
+        """
+        if not ref:
+            raise ValueError("trodo: export_dataset(ref) requires a dataset ref")
+        if not token:
+            raise ValueError("trodo: export_dataset requires an export token (Settings → Datasets)")
+        params: Dict[str, Any] = {"labels": labels}
+        if include_unlabelled:
+            params["include_unlabelled"] = 1
+        if version is not None:
+            params["version"] = int(version)
+        if since:
+            params["since"] = since
+        res = self._http.export_dataset(ref, params, token)
+        if res.get("__error"):
+            raise RuntimeError(
+                f"trodo: export of {ref!r} failed: HTTP {res.get('status')} {res.get('message') or res.get('error') or ''}".strip()
+            )
+        rows: List[Dict[str, Any]] = []
+        for line in (res.get("text") or "").split("\n"):
+            line = line.strip()
+            if line:
+                rows.append(json.loads(line))
+        return {"rows": rows, "version": res.get("version")}

{trodo_python-2.25.0.dist-info → trodo_python-2.26.1.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: trodo-python
-Version: 2.25.0
+Version: 2.26.1
 Summary: Trodo Analytics SDK for Python — server-side event tracking
 License: ISC
 Keywords: analytics,tracking,trodo,server-side

{trodo_python-2.25.0.dist-info → trodo_python-2.26.1.dist-info}/RECORD

@@ -1,5 +1,5 @@
-trodo/__init__.py,sha256=ftnDcQvvOi9VnJoynkzzZCNwqHRhC7pbXS836bTwizo,27913
-trodo/_version.py,sha256=50fFgmf3lAeDxqh-LvVujTqoWvZoG7fTVAY5gKvc-yQ,23
+trodo/__init__.py,sha256=RCoNO_seZkIUmweHvyKb8t6J5ydktoAHOg2nt2sp448,28752
+trodo/_version.py,sha256=MX7IGM5dYbXPeFIXOWPhV7DPbCslhEe8Q7DCvxwKmhY,23
 trodo/client.py,sha256=6k6HeKrGV2720kK17--u6BjbhAUFZNUzpd9yEYiaQOs,24137
 trodo/types.py,sha256=eySgUvCXROG2TxtxgiU0MNr5iH0DEcduK8bmYtTKG44,3138
 trodo/user_context.py,sha256=uHCI2WYoOI3cNwdIUEuZdX4VYSu14pzm348ksn6hn_c,8195
@@ -7,13 +7,13 @@ trodo/api/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 trodo/api/_retry.py,sha256=mHhuzfe3QuYa6j9BqRYVClvnYH0o3WkXXngmSwqWfgY,1214
 trodo/api/async_client.py,sha256=qF97NbvoYbNhem8uXtPh14EG4JVAhEAzUCYn7busaaA,3609
 trodo/api/endpoints.py,sha256=SGgVELcVfwuCrRwlVPW2j0KYkaMtwLFPpdMtGtlqPeA,1574
-trodo/api/http_client.py,sha256=Vnw8l_QyoslCF2lJIxoH2uoUNFAK-8S_YfnWN8f4ZS0,11362
+trodo/api/http_client.py,sha256=iOG9D4ackVAz0nQD5ZjwfIvRiEmgxga9uXgxrRQX6L4,13185
 trodo/auto/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 trodo/auto/auto_event_manager.py,sha256=cztuRsRkNoJE5R4NfSfTrTJTGl4jx2Yb-Ncy0aVAPo8,4247
 trodo/delivery/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 trodo/delivery/sender.py,sha256=urO7m6mRMnSEW7_cNAmJ4A6jtDNxuihZHOxCx1jQScI,21348
 trodo/managers/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-trodo/managers/dataset_manager.py,sha256=gx0S8ujG3cbw4IskBNMdceBjDbcEh_nebwKSzEvm7hE,2827
+trodo/managers/dataset_manager.py,sha256=mifHYkHLWo0olXCNfQOqvpLPL8DDgbub7YAWWlrjIHQ,5609
 trodo/managers/experiment_manager.py,sha256=V-vemLtenfSI5VIcP0HPiXzEy8AyTrRrbJVCA9Z6wjo,10579
 trodo/managers/group_manager.py,sha256=ki3Se3qEoSZfREX63oeDeBmEfZF-ISHLE8azEtLg0tM,3542
 trodo/managers/people_manager.py,sha256=mMVnx40Mlifx6NGgChvohC9ViK6dQu2mkXNHbV8pK1E,2882
@@ -40,7 +40,7 @@ trodo/session/server_session.py,sha256=McsudEiq33XDq3nqxgzBcUvIjQxCMscwEuAPnYXrT
 trodo/session/session_manager.py,sha256=7ht5LeeZX1HLLfPeNV_a8NkXGbHl3up0CRtCGr3EzjQ,2995
 trodo/util/__init__.py,sha256=Z9c4rPPdKg06Kk3byKheDqksSgR4WNvS475oQ5sNljc,54
 trodo/util/lru.py,sha256=QIsM7s6J_E9ZjiXatSyReZIEeGypTAcgrPWWy8YsRa4,2465
-trodo_python-2.25.0.dist-info/METADATA,sha256=_e2LmqFYjaZQdhh6Nr-Uj9wr_f88ZYLZDlD17N431as,25308
-trodo_python-2.25.0.dist-info/WHEEL,sha256=YVMoNqKzERt-wjUZwJ33xBGAwnFl-4cqbYkTtWa4itE,91
-trodo_python-2.25.0.dist-info/top_level.txt,sha256=VCQu1CJWFmNsqTs1YxMcw4Mq35Tc3z3uI9RwHEXAayQ,6
-trodo_python-2.25.0.dist-info/RECORD,,
+trodo_python-2.26.1.dist-info/METADATA,sha256=2I7mNSS6gC_w6I3Mc9iOT-E6VHLboXBB9o8i4F-8YN0,25308
+trodo_python-2.26.1.dist-info/WHEEL,sha256=YVMoNqKzERt-wjUZwJ33xBGAwnFl-4cqbYkTtWa4itE,91
+trodo_python-2.26.1.dist-info/top_level.txt,sha256=VCQu1CJWFmNsqTs1YxMcw4Mq35Tc3z3uI9RwHEXAayQ,6
+trodo_python-2.26.1.dist-info/RECORD,,