inferencesh 0.3.1__py3-none-any.whl → 0.4.0__py3-none-any.whl

inferencesh/models/file.py

@@ -5,11 +5,45 @@ import os
 import urllib.request
 import urllib.parse
 import tempfile
+import hashlib
+from pathlib import Path
 from tqdm import tqdm
 
 
 class File(BaseModel):
     """A class representing a file in the inference.sh ecosystem."""
+    
+    @classmethod
+    def get_cache_dir(cls) -> Path:
+        """Get the cache directory path based on environment variables or default location."""
+        if cache_dir := os.environ.get("FILE_CACHE_DIR"):
+            path = Path(cache_dir)
+        else:
+            path = Path.home() / ".cache" / "inferencesh" / "files"
+        path.mkdir(parents=True, exist_ok=True)
+        return path
+    
+    def _get_cache_path(self, url: str) -> Path:
+        """Get the cache path for a URL using a hash-based directory structure."""
+        # Parse URL components
+        parsed_url = urllib.parse.urlparse(url)
+        
+        # Create hash from URL path and query parameters for uniqueness
+        url_components = parsed_url.netloc + parsed_url.path
+        if parsed_url.query:
+            url_components += '?' + parsed_url.query
+        url_hash = hashlib.sha256(url_components.encode()).hexdigest()[:12]
+        
+        # Get filename from URL or use default
+        filename = os.path.basename(parsed_url.path)
+        if not filename:
+            filename = 'download'
+            
+        # Create hash directory in cache
+        cache_dir = self.get_cache_dir() / url_hash
+        cache_dir.mkdir(exist_ok=True)
+        
+        return cache_dir / filename
     uri: Optional[str] = Field(default=None)  # Original location (URL or file path)
     path: Optional[str] = None  # Resolved local file path
     content_type: Optional[str] = None  # MIME type of the file
@@ -74,11 +108,20 @@ class File(BaseModel):
         return parsed.scheme in ('http', 'https')
 
     def _download_url(self) -> None:
-        """Download the URL to a temporary file and update the path."""
+        """Download the URL to the cache directory and update the path."""
         original_url = self.uri
+        cache_path = self._get_cache_path(original_url)
+        
+        # If file exists in cache, use it
+        if cache_path.exists():
+            print(f"Using cached file: {cache_path}")
+            self.path = str(cache_path)
+            return
+            
+        print(f"Downloading URL: {original_url} to {cache_path}")
         tmp_file = None
         try:
-            # Create a temporary file with a suffix based on the URL path
+            # Download to temporary file first to avoid partial downloads in cache
             suffix = os.path.splitext(urllib.parse.urlparse(original_url).path)[1]
             tmp_file = tempfile.NamedTemporaryFile(delete=False, suffix=suffix)
             self._tmp_path = tmp_file.name
@@ -133,7 +176,10 @@ class File(BaseModel):
                                     # If we read the whole body at once, exit loop
                                     break
                             
-                self.path = self._tmp_path
+                # Move the temporary file to the cache location
+                os.replace(self._tmp_path, cache_path)
+                self._tmp_path = None  # Prevent deletion in __del__
+                self.path = str(cache_path)
             except (urllib.error.URLError, urllib.error.HTTPError) as e:
                 raise RuntimeError(f"Failed to download URL {original_url}: {str(e)}")
             except IOError as e:

inferencesh/utils/download.py

@@ -24,16 +24,24 @@ def download(url: str, directory: Union[str, Path, StorageDir]) -> str:
     dir_path = Path(directory)
     dir_path.mkdir(exist_ok=True)
     
-    # Create hash directory from URL
-    url_hash = hashlib.sha256(url.encode()).hexdigest()[:12]
-    hash_dir = dir_path / url_hash
-    hash_dir.mkdir(exist_ok=True)
+    # Parse URL components
+    parsed_url = urllib.parse.urlparse(url)
     
-    # Keep original filename
-    filename = os.path.basename(urllib.parse.urlparse(url).path)
+    # Create hash from URL path and query parameters for uniqueness
+    url_components = parsed_url.netloc + parsed_url.path
+    if parsed_url.query:
+        url_components += '?' + parsed_url.query
+    url_hash = hashlib.sha256(url_components.encode()).hexdigest()[:12]
+    
+    # Keep original filename or use a default
+    filename = os.path.basename(parsed_url.path)
     if not filename:
         filename = 'download'
-        
+    
+    # Create hash directory and store file
+    hash_dir = dir_path / url_hash
+    hash_dir.mkdir(exist_ok=True)
+    
     output_path = hash_dir / filename
     
     # If file exists in directory and it's not a temp directory, return it

{inferencesh-0.3.1.dist-info → inferencesh-0.4.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: inferencesh
-Version: 0.3.1
+Version: 0.4.0
 Summary: inference.sh Python SDK
 Author: Inference Shell Inc.
 Author-email: "Inference Shell Inc." <hello@inference.sh>

{inferencesh-0.3.1.dist-info → inferencesh-0.4.0.dist-info}/RECORD

@@ -2,14 +2,14 @@ inferencesh/__init__.py,sha256=dY3l3yCkWoMtGX0gNXgxFnrprFRl6PPWjH8V7Qedx5g,772
 inferencesh/client.py,sha256=cm7E-8LxP8jyb7JnANmcBtF1ya2i7sMBW2Pq-Oh-mcM,31318
 inferencesh/models/__init__.py,sha256=FDwcdtT6c4hbRitymjmN-hZMlQa8RbKSftkZZyjtUXA,536
 inferencesh/models/base.py,sha256=4gZQRi8J7y9U6PrGD9pRIehd1MJVJAqGakPQDs2AKFM,3251
-inferencesh/models/file.py,sha256=0CSbIoFTvGT1CmGnMjkNmGR1N-zg64bmheVUWiCczxE,8714
+inferencesh/models/file.py,sha256=uh1czgk0KFl_9RHTODX0PkdnI42MSU8QMJR_I4lVKI4,10556
 inferencesh/models/llm.py,sha256=GLcEkDizBbgcfc-zC719wDe44th3EGf3FpKERjIAPE8,27755
 inferencesh/utils/__init__.py,sha256=-xiD6uo2XzcrPAWFb_fUbaimmnW4KFKc-8IvBzaxNd4,148
-inferencesh/utils/download.py,sha256=7n5twvoNYDcFnKJyefImaj2YfzRI7vddQw4usZbj38c,1521
+inferencesh/utils/download.py,sha256=DRGBudiPVa5bDS35KfR-DYeGRk7gO03WOelnisecwMo,1815
 inferencesh/utils/storage.py,sha256=E4J8emd4eFKdmdDgAqzz3TpaaDd3n0l8gYlMHuY8yIU,519
-inferencesh-0.3.1.dist-info/licenses/LICENSE,sha256=OsgqEWIh2el_QMj0y8O1A5Q5Dl-dxqqYbFE6fszuR4s,1086
-inferencesh-0.3.1.dist-info/METADATA,sha256=gU67lD_F5c66kAhn375nGfLpP57ueG5PPrlJsvPD8uY,2964
-inferencesh-0.3.1.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-inferencesh-0.3.1.dist-info/entry_points.txt,sha256=6IC-fyozAqW3ljsMLGCXxJ0_ui2Jb-2fLHtoH1RTnEE,45
-inferencesh-0.3.1.dist-info/top_level.txt,sha256=TSMHg3T1ThMl1HGAWmzBClwOYH1ump5neof9BfHIwaA,12
-inferencesh-0.3.1.dist-info/RECORD,,
+inferencesh-0.4.0.dist-info/licenses/LICENSE,sha256=OsgqEWIh2el_QMj0y8O1A5Q5Dl-dxqqYbFE6fszuR4s,1086
+inferencesh-0.4.0.dist-info/METADATA,sha256=pHnblJABrxy5Iy81hpP7nV-J72Tp2JIUJ6D2UzVbSqo,2964
+inferencesh-0.4.0.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+inferencesh-0.4.0.dist-info/entry_points.txt,sha256=6IC-fyozAqW3ljsMLGCXxJ0_ui2Jb-2fLHtoH1RTnEE,45
+inferencesh-0.4.0.dist-info/top_level.txt,sha256=TSMHg3T1ThMl1HGAWmzBClwOYH1ump5neof9BfHIwaA,12
+inferencesh-0.4.0.dist-info/RECORD,,