sutro 0.1.27__py3-none-any.whl → 0.1.28__py3-none-any.whl

sutro/cli.py

@@ -141,7 +141,6 @@ def jobs():
     """Manage jobs."""
     pass
 
-
 @jobs.command()
 @click.option(
     "--all", is_flag=True, help="Include all jobs, including cancelled and failed ones."
@@ -360,6 +359,24 @@ def download(dataset_id, file_name=None, output_path=None):
             with open(output_path + "/" + file_name, "wb") as f:
                 f.write(file)
 
+@cli.group()
+def cache():
+    """Manage the local job results cache."""
+    pass
+
+@cache.command()
+def clear():
+    """Clear the local job results cache."""
+    sdk = get_sdk()
+    sdk._clear_job_results_cache()
+    click.echo(Fore.GREEN + "Job results cache cleared." + Style.RESET_ALL)
+
+@cache.command()
+def show():
+    """Show the contents and size of the job results cache."""
+    sdk = get_sdk()
+    sdk._show_cache_contents()
+
 
 @cli.command()
 def docs():
@@ -395,6 +412,7 @@ def quotas():
         + Style.RESET_ALL
     )
 
+
 @jobs.command()
 @click.argument("job_id", required=False)
 @click.option("--latest", is_flag=True, help="Attach to the latest job.")

sutro/sdk.py

@@ -17,6 +17,8 @@ from tqdm import tqdm
 import time
 from pydantic import BaseModel
 import json
+import pyarrow.parquet as pq
+import shutil
 
 
 class JobStatus(str, Enum):
@@ -819,6 +821,7 @@ class Sutro:
             include_cumulative_logprobs: bool = False,
             with_original_df: pl.DataFrame | pd.DataFrame = None,
             output_column: str = "inference_result",
+            disable_cache: bool = False
     ):
         """
         Get the results of a job by its ID.
@@ -835,43 +838,66 @@ class Sutro:
         Returns:
             Union[pl.DataFrame, pd.DataFrame]: The results as a DataFrame. By default, returns polars.DataFrame; when with_original_df is an instance of pandas.DataFrame, returns pandas.DataFrame.
         """
-        endpoint = f"{self.base_url}/job-results"
-        payload = {
-            "job_id": job_id,
-            "include_inputs": include_inputs,
-            "include_cumulative_logprobs": include_cumulative_logprobs,
-        }
-        headers = {
-            "Authorization": f"Key {self.api_key}",
-            "Content-Type": "application/json",
-        }
-        with yaspin(
+
+        
+        file_path = os.path.expanduser(f"~/.sutro/job-results/{job_id}.snappy.parquet")
+        expected_num_columns = 1 + include_inputs + include_cumulative_logprobs
+        contains_expected_columns = False
+        if os.path.exists(file_path):
+            num_columns = pq.read_table(file_path).num_columns
+            contains_expected_columns = num_columns == expected_num_columns
+        
+        if disable_cache == False and contains_expected_columns:
+            with yaspin(
+                SPINNER,
+                text=to_colored_text(f"Loading results from cache: {file_path}"),
+                color=YASPIN_COLOR,
+            ) as spinner:
+                results_df = pl.read_parquet(file_path)
+                spinner.write(to_colored_text("✔ Results loaded from cache", state="success"))
+        else:
+            endpoint = f"{self.base_url}/job-results"
+            payload = {
+                "job_id": job_id,
+                "include_inputs": include_inputs,
+                "include_cumulative_logprobs": include_cumulative_logprobs,
+            }
+            headers = {
+                "Authorization": f"Key {self.api_key}",
+                "Content-Type": "application/json",
+            }
+            with yaspin(
                 SPINNER,
                 text=to_colored_text(f"Gathering results from job: {job_id}"),
                 color=YASPIN_COLOR,
-        ) as spinner:
-            response = requests.post(
-                endpoint, data=json.dumps(payload), headers=headers
-            )
-            if response.status_code != 200:
-                spinner.write(
-                    to_colored_text(
-                        f"Bad status code: {response.status_code}", state="fail"
-                    )
+            ) as spinner:
+                response = requests.post(
+                    endpoint, data=json.dumps(payload), headers=headers
                 )
-                spinner.stop()
-                print(to_colored_text(response.json(), state="fail"))
-                return None
+                if response.status_code != 200:
+                    spinner.write(
+                        to_colored_text(
+                            f"Bad status code: {response.status_code}", state="fail"
+                        )
+                    )
+                    spinner.stop()
+                    print(to_colored_text(response.json(), state="fail"))
+                    return None
 
-            spinner.write(
-                to_colored_text("✔ Job results retrieved", state="success")
-            )
+                spinner.write(
+                    to_colored_text("✔ Job results retrieved", state="success")
+                )
 
-        response_data = response.json()
-        results_df = pl.DataFrame(response_data["results"])
+            response_data = response.json()
+            results_df = pl.DataFrame(response_data["results"])
 
-        results_df = results_df.rename({'outputs': output_column})
+            results_df = results_df.rename({'outputs': output_column})
 
+            if disable_cache == False:
+                os.makedirs(os.path.dirname(file_path), exist_ok=True)
+                results_df.write_parquet(file_path, compression="snappy")
+                spinner.write(to_colored_text("✔ Results saved to cache", state="success"))
+                
         # Ordering inputs col first seems most logical/useful
         column_config = [
             ('inputs', include_inputs),
@@ -1296,6 +1322,31 @@ class Sutro:
                 time.sleep(POLL_INTERVAL)
 
         return results
+
+    def _clear_job_results_cache(self): # only to be called by the CLI
+        """
+        Clears the cache for a job results.
+        """
+        if os.path.exists(os.path.expanduser("~/.sutro/job-results")):
+            shutil.rmtree(os.path.expanduser("~/.sutro/job-results"))
+
+    def _show_cache_contents(self):
+        """
+        Shows the contents and size of each file in the job results cache.
+        """
+        # get the size of the job-results directory
+        with yaspin(
+            SPINNER, text=to_colored_text("Retrieving job results cache contents"), color=YASPIN_COLOR
+        ) as spinner:
+            if not os.path.exists(os.path.expanduser("~/.sutro/job-results")):
+                spinner.write(to_colored_text("No job results cache found", "success"))
+                return
+            total_size = 0
+            for file in os.listdir(os.path.expanduser("~/.sutro/job-results")):
+                size = os.path.getsize(os.path.expanduser(f"~/.sutro/job-results/{file}")) / 1024 / 1024 / 1024
+                total_size += size
+                spinner.write(to_colored_text(f"File: {file} - Size: {size} GB"))
+            spinner.write(to_colored_text(f"Total size of results cache at ~/.sutro/job-results: {total_size} GB", "success"))
     
     def _await_job_start(self, job_id: str, timeout: Optional[int] = 7200):
         """

{sutro-0.1.27.dist-info → sutro-0.1.28.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: sutro
-Version: 0.1.27
+Version: 0.1.28
 Summary: Sutro Python SDK
 Project-URL: Homepage, https://sutro.sh
 Project-URL: Documentation, https://docs.sutro.sh
@@ -12,6 +12,7 @@ Requires-Dist: colorama==0.4.4
 Requires-Dist: numpy==2.1.1
 Requires-Dist: pandas==2.2.3
 Requires-Dist: polars==1.8.2
+Requires-Dist: pyarrow==21.0.0
 Requires-Dist: pydantic==2.11.4
 Requires-Dist: requests==2.32.3
 Requires-Dist: tqdm==4.67.1

sutro-0.1.28.dist-info/RECORD

@@ -0,0 +1,8 @@
+sutro/__init__.py,sha256=yUiVwcZ8QamSqDdRHgzoANyTZ-x3cPzlt2Fs5OllR_w,402
+sutro/cli.py,sha256=8DrJVbjoayCUz4iszlj35Tv1q1gzDVzx_CuF6gZHwuU,13636
+sutro/sdk.py,sha256=0_KC8eOOEunEPChrpToIucbg6q-icyxSUeMwd1UKDSY,54150
+sutro-0.1.28.dist-info/METADATA,sha256=Nu730OdGc5TMVMlw3_kqxZ03z8_3jsz57Og5PIhk7Wo,700
+sutro-0.1.28.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
+sutro-0.1.28.dist-info/entry_points.txt,sha256=eXvr4dvMV4UmZgR0zmrY8KOmNpo64cJkhNDywiadRFM,40
+sutro-0.1.28.dist-info/licenses/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
+sutro-0.1.28.dist-info/RECORD,,

sutro-0.1.27.dist-info/RECORD

@@ -1,8 +0,0 @@
-sutro/__init__.py,sha256=yUiVwcZ8QamSqDdRHgzoANyTZ-x3cPzlt2Fs5OllR_w,402
-sutro/cli.py,sha256=6Qy9Vwaaho92HeO8YA_z1De4zp1dEFkSX3bEnLvdbkE,13203
-sutro/sdk.py,sha256=zBnLIg9wvoAplGEVNdgfCsgnUKR48NDPVOB4rBSYR4M,51552
-sutro-0.1.27.dist-info/METADATA,sha256=KO6cMAG83DpMnP0wf9B_9Fda1_zipO9Gqnfy-kn5HuM,669
-sutro-0.1.27.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
-sutro-0.1.27.dist-info/entry_points.txt,sha256=eXvr4dvMV4UmZgR0zmrY8KOmNpo64cJkhNDywiadRFM,40
-sutro-0.1.27.dist-info/licenses/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
-sutro-0.1.27.dist-info/RECORD,,