attachments 0.2.3__tar.gz → 0.2.6__tar.gz

{attachments-0.2.3/src/attachments.egg-info → attachments-0.2.6}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: attachments
-Version: 0.2.3
+Version: 0.2.6
 Summary: A Python library to handle various file types for LLMs.
 Author-email: Maxim Rivest <mrive052@gmail.com>
 License-Expression: MIT
@@ -35,13 +35,16 @@ Requires-Dist: flake8; extra == "dev"
 Requires-Dist: mypy; extra == "dev"
 Requires-Dist: black; extra == "dev"
 Requires-Dist: filelock; extra == "dev"
+Requires-Dist: mystmd; extra == "dev"
 Dynamic: license-file
 
-# Attachments
+# Attachments: The Python Library for Effortless LLM Context
 ## Easiest way to give context to LLMs
 
 Attachments has the ambition to be the general funnel for any files to be transformed into images+text for large language models context. For this to be true we, has a community, have to stop implementing this funnel (file -> images+text) each on our own. If you have a detector + parser + renderer contibute it! This will help you and us. Don't let your code alone in a dark corner of you computer.
 
+From a `path` to a `string` and a `base64` image encoded string. That is the goal.
+
 ## Very quickstart
 
 ```bash
@@ -121,6 +124,43 @@ message = anthropic.Anthropic().messages.create(
 print(message.content)
 ```
 
+## How to give a word document (with tables and images) to langchain?
+
+I am not a fan of the object that langchain and anthropic make you create, but hey!, I don't make the rules...
+
+```python
+from langchain.chat_models import init_chat_model
+from attachments import Attachments
+
+docx_file = Attachments("https://github.com/microsoft/markitdown/raw/refs/heads/main/packages/markitdown/tests/test_files/test.docx")
+
+llm = init_chat_model("anthropic:claude-3-5-sonnet-latest")
+
+# Creation of a list of object as expected by langchain/anthropic
+content = [{"type": "text", "text": f"Analyze the following documents: {docx_file}"}]
+
+for image in docx_file.images
+    media_type_i = image.split(";")[0].split(":")[1]
+    media_data_i = image.split(",")[1]
+    content.append({
+        "type": "image",
+        "source": {
+            "type": "base64",
+            "media_type": media_type_i, 
+            "data": media_data_i 
+        }
+    })
+
+
+message = {
+    "role": "user",
+    "content": content,
+}
+
+#Finally call the llm
+response = llm.invoke([message])
+print(response.text())
+```
 
 ## Usage
 

{attachments-0.2.3 → attachments-0.2.6}/README.md

@@ -1,8 +1,10 @@
-# Attachments
+# Attachments: The Python Library for Effortless LLM Context
 ## Easiest way to give context to LLMs
 
 Attachments has the ambition to be the general funnel for any files to be transformed into images+text for large language models context. For this to be true we, has a community, have to stop implementing this funnel (file -> images+text) each on our own. If you have a detector + parser + renderer contibute it! This will help you and us. Don't let your code alone in a dark corner of you computer.
 
+From a `path` to a `string` and a `base64` image encoded string. That is the goal.
+
 ## Very quickstart
 
 ```bash
@@ -82,6 +84,43 @@ message = anthropic.Anthropic().messages.create(
 print(message.content)
 ```
 
+## How to give a word document (with tables and images) to langchain?
+
+I am not a fan of the object that langchain and anthropic make you create, but hey!, I don't make the rules...
+
+```python
+from langchain.chat_models import init_chat_model
+from attachments import Attachments
+
+docx_file = Attachments("https://github.com/microsoft/markitdown/raw/refs/heads/main/packages/markitdown/tests/test_files/test.docx")
+
+llm = init_chat_model("anthropic:claude-3-5-sonnet-latest")
+
+# Creation of a list of object as expected by langchain/anthropic
+content = [{"type": "text", "text": f"Analyze the following documents: {docx_file}"}]
+
+for image in docx_file.images
+    media_type_i = image.split(";")[0].split(":")[1]
+    media_data_i = image.split(",")[1]
+    content.append({
+        "type": "image",
+        "source": {
+            "type": "base64",
+            "media_type": media_type_i, 
+            "data": media_data_i 
+        }
+    })
+
+
+message = {
+    "role": "user",
+    "content": content,
+}
+
+#Finally call the llm
+response = llm.invoke([message])
+print(response.text())
+```
 
 ## Usage
 

{attachments-0.2.3 → attachments-0.2.6}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "attachments"
-version = "0.2.3"
+version = "0.2.6"
 authors = [
   { name="Maxim Rivest", email="mrive052@gmail.com" },
 ]
@@ -53,6 +53,7 @@ dev = [
     "mypy",
     "black",
     "filelock",
+    "mystmd",
 ]
 
 [tool.setuptools]

{attachments-0.2.3 → attachments-0.2.6}/src/attachments/core.py

@@ -278,7 +278,7 @@ class Attachments:
                         'm4a': 'audio/m4a', # mimetypes might say audio/mp4, which is also fine for .m4a
                         'mp4_audio': 'audio/mp4',
                         'wav': 'audio/wav', 
-                        'flac': 'audio/flac', 
+                        'flac': 'audio/flac',
                         'ogg_audio': 'audio/ogg',
                         'oga': 'audio/ogg', 
                         'webm_audio': 'audio/webm', 
@@ -634,6 +634,59 @@ class Attachments:
     def another_method_or_end_of_class_placeholder(self):
         pass
 
+    def to_openai_content(self, prompt: str) -> List[Dict[str, Any]]:
+        """
+        Formats attachments and a prompt for the OpenAI API.
+        Currently supports images.
+
+        Args:
+            prompt: The text prompt to include.
+
+        Returns:
+            A list of dictionaries formatted for the OpenAI API,
+            combining image data and the text prompt.
+        """
+        content = [{"type": "input_image", "image_url": image_data_uri} for image_data_uri in self.images]
+        content.append({"type": "input_text", "text": prompt + "\n\n" + self.__str__()})
+        return content
+
+    def to_claude_content(self, prompt: str) -> List[Dict[str, Any]]:
+        """
+        Formats attachments and a prompt for the Anthropic Claude API.
+        Currently supports images.
+
+        Args:
+            prompt: The text prompt to include.
+
+        Returns:
+            A list of dictionaries formatted for the Anthropic Claude API,
+            combining image data and the text prompt.
+        """
+        content = []
+        for image_data_uri in self.images:
+            try:
+                # Example data URI: "data:image/jpeg;base64,BASE64_STRING"
+                header, base64_data = image_data_uri.split(',', 1)
+                media_type = header.split(';')[0].split(':')[1]
+                
+                content.append({
+                    "type": "image",
+                    "source": {
+                        "type": "base64",
+                        "media_type": media_type,
+                        "data": base64_data
+                    }
+                })
+            except ValueError as e:
+                # Handle cases where the data URI might not be as expected, though self.images should provide valid ones.
+                if self.verbose:
+                    print(f"Warning: Could not parse image data URI for Claude content: {image_data_uri}. Error: {e}")
+                # Optionally, append an error or skip this image
+                continue # Skip malformed URIs for now
+        
+        content.append({"type": "text", "text": prompt})
+        return content
+
     # Potentially other methods follow, or end of class / file 
 
     def _format_operations_for_markdown(self, operations: Dict[str, Any]) -> str:

{attachments-0.2.3 → attachments-0.2.6/src/attachments.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: attachments
-Version: 0.2.3
+Version: 0.2.6
 Summary: A Python library to handle various file types for LLMs.
 Author-email: Maxim Rivest <mrive052@gmail.com>
 License-Expression: MIT
@@ -35,13 +35,16 @@ Requires-Dist: flake8; extra == "dev"
 Requires-Dist: mypy; extra == "dev"
 Requires-Dist: black; extra == "dev"
 Requires-Dist: filelock; extra == "dev"
+Requires-Dist: mystmd; extra == "dev"
 Dynamic: license-file
 
-# Attachments
+# Attachments: The Python Library for Effortless LLM Context
 ## Easiest way to give context to LLMs
 
 Attachments has the ambition to be the general funnel for any files to be transformed into images+text for large language models context. For this to be true we, has a community, have to stop implementing this funnel (file -> images+text) each on our own. If you have a detector + parser + renderer contibute it! This will help you and us. Don't let your code alone in a dark corner of you computer.
 
+From a `path` to a `string` and a `base64` image encoded string. That is the goal.
+
 ## Very quickstart
 
 ```bash
@@ -121,6 +124,43 @@ message = anthropic.Anthropic().messages.create(
 print(message.content)
 ```
 
+## How to give a word document (with tables and images) to langchain?
+
+I am not a fan of the object that langchain and anthropic make you create, but hey!, I don't make the rules...
+
+```python
+from langchain.chat_models import init_chat_model
+from attachments import Attachments
+
+docx_file = Attachments("https://github.com/microsoft/markitdown/raw/refs/heads/main/packages/markitdown/tests/test_files/test.docx")
+
+llm = init_chat_model("anthropic:claude-3-5-sonnet-latest")
+
+# Creation of a list of object as expected by langchain/anthropic
+content = [{"type": "text", "text": f"Analyze the following documents: {docx_file}"}]
+
+for image in docx_file.images
+    media_type_i = image.split(";")[0].split(":")[1]
+    media_data_i = image.split(",")[1]
+    content.append({
+        "type": "image",
+        "source": {
+            "type": "base64",
+            "media_type": media_type_i, 
+            "data": media_data_i 
+        }
+    })
+
+
+message = {
+    "role": "user",
+    "content": content,
+}
+
+#Finally call the llm
+response = llm.invoke([message])
+print(response.text())
+```
 
 ## Usage
 

{attachments-0.2.3 → attachments-0.2.6}/src/attachments.egg-info/requires.txt

@@ -23,3 +23,4 @@ flake8
 mypy
 black
 filelock
+mystmd