Skip to content

Commit

Permalink
Merge pull request #16 from Zipstack/url-corrections
Browse files Browse the repository at this point in the history 
feat: Base url changes and README.md changes
  • Loading branch information
@arun-venkataswamy
arun-venkataswamy authored Nov 5, 2024
2 parents 0075b70 + abaf736 commit 82baf42
Show file tree
Hide file tree
Showing 3 changed files with 30 additions and 133 deletions.
129 changes: 6 additions & 123 deletions README.md
View file
Original file line number Diff line number Diff line change
Expand Up @@ -7,135 +7,18 @@

LLMs are powerful, but their output is as good as the input you provide. LLMWhisperer is a technology that presents data from complex documents (different designs and formats) to LLMs in a way that they can best understand. LLMWhisperer features include Layout Preserving Mode, Auto-switching between native text and OCR modes, proper representation of radio buttons and checkboxes in PDF forms as raw text, among other features. You can now extract raw text from complex PDF documents or images without having to worry about whether the document is a native text document, a scanned image or just a picture clicked on a smartphone. Extraction of raw text from invoices, purchase orders, bank statements, etc works easily for structured data extraction with LLMs powered by LLMWhisperer's Layout Preserving mode.

Refer to the client documentation for more information: [LLMWhisperer Client Documentation](https://docs.unstract.com/llmwhisperer/index.html)
Refer to the client documentation for more information: [LLMWhisperer Client Documentation](https://docs.unstract.com/llmwhisperer/llm_whisperer/python_client/llm_whisperer_python_client_intro/)

## Features
## A note on versions

- Easy to use Pythonic interface.
- Handles all the HTTP requests and responses for you.
- Raises Python exceptions for API errors.
There are two versions of the client library available in this package:

## Installation
**LLMWhispererClient**: This is the legacy version of the client library and is recommended for supporting older apps only. This version will be deprecated in the future.

You can install the LLMWhisperer Python Client using pip:
**LLMWhispererClientV2**: This is the latest version of the client library and is recommended for all new users. It is mandatory for all users who are using LLMWhisperer API version 2.0.0 and above (All customers who have signed up after 5th November 2024).

```bash
pip install llmwhisperer-client
```
Documentation for both versions are available [here](https://docs.unstract.com/llmwhisperer/)

## Usage

First, import the `LLMWhispererClient` from the `client` module:

```python
from unstract.llmwhisperer.client import LLMWhispererClient
```

Then, create an instance of the `LLMWhispererClient`:

```python
client = LLMWhispererClient(base_url="https://llmwhisperer-api.unstract.com/v1", api_key="your_api_key")
```

Now, you can use the client to interact with the LLMWhisperer API:

```python
# Get usage info
usage_info = client.get_usage_info()

# Process a document
# Extracted text is available in the 'extracted_text' field of the result
whisper = client.whisper(file_path="path_to_your_file")

# Get the status of a whisper operation
# whisper_hash is available in the 'whisper_hash' field of the result of the whisper operation
status = client.whisper_status(whisper_hash)

# Retrieve the result of a whisper operation
# whisper_hash is available in the 'whisper_hash' field of the result of the whisper operation
whisper = client.whisper_retrieve(whisper_hash)
```

### Error Handling

The client raises `LLMWhispererClientException` for API errors:

```python
try:
result = client.whisper_retrieve("invalid_hash")
except LLMWhispererClientException as e:
print(f"Error: {e.message}, Status Code: {e.status_code}")
```

### Simple use case with defaults

```python
client = LLMWhispererClient()
try:
result = client.whisper(file_path="sample_files/restaurant_invoice_photo.pdf")
extracted_text = result["extracted_text"]
print(extracted_text)
except LLMWhispererClientException as e:
print(e)
```

### Simple use case with more options set
We are forcing text processing and extracting text from the first two pages only.

```python
client = LLMWhispererClient()
try:
result = client.whisper(
file_path="sample_files/credit_card.pdf",
processing_mode="text",
force_text_processing=True,
pages_to_extract="1,2",
)
extracted_text = result["extracted_text"]
print(extracted_text)
except LLMWhispererClientException as e:
print(e)
```

### Extraction with timeout set

The platform has a hard timeout of 200 seconds. If the document takes more than 200 seconds to convert (large documents), the platform will switch to async extraction and return a hash. The client can be used to check the status of the extraction and retrieve the result. Also note that the timeout is in seconds and can be set by the caller too.


```python
client = LLMWhispererClient()
try:
result = client.whisper(
file_path="sample_files/credit_card.pdf",
pages_to_extract="1,2",
timeout=2,
)
if result["status_code"] == 202:
print("Timeout occured. Whisper request accepted.")
print(f"Whisper hash: {result['whisper-hash']}")
while True:
print("Polling for whisper status...")
status = client.whisper_status(whisper_hash=result["whisper-hash"])
if status["status"] == "processing":
print("STATUS: processing...")
elif status["status"] == "delivered":
print("STATUS: Already delivered!")
break
elif status["status"] == "unknown":
print("STATUS: unknown...")
break
elif status["status"] == "processed":
print("STATUS: processed!")
print("Let's retrieve the result of the extraction...")
resultx = client.whisper_retrieve(
whisper_hash=result["whisper-hash"]
)
print(resultx["extracted_text"])
break
time.sleep(2)
except LLMWhispererClientException as e:
print(e)
```

## Questions and Feedback

Expand Down
2 changes: 1 addition & 1 deletion src/unstract/llmwhisperer/__init__.py
View file
Original file line number Diff line number Diff line change
@@ -1,4 +1,4 @@
__version__ = "0.23.0"
__version__ = "2.0.0"

from .client import LLMWhispererClient # noqa: F401
from .client_v2 import LLMWhispererClientV2 # noqa: F401
Expand Down
32 changes: 23 additions & 9 deletions src/unstract/llmwhisperer/client_v2.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -27,7 +27,7 @@

import requests

BASE_URL = "https://llmwhisperer-api.unstract.com/api/v2"
BASE_URL_V2 = "https://llmwhisperer-api.us-central.unstract.com/api/v2"


class LLMWhispererClientException(Exception):
Expand Down Expand Up @@ -62,7 +62,9 @@ class LLMWhispererClientV2:
client's activities and errors.
"""

formatter = logging.Formatter("%(asctime)s - %(name)s - %(levelname)s - %(message)s")
formatter = logging.Formatter(
"%(asctime)s - %(name)s - %(levelname)s - %(message)s"
)
logger = logging.getLogger(__name__)
log_stream_handler = logging.StreamHandler()
log_stream_handler.setFormatter(formatter)
Expand Down Expand Up @@ -108,7 +110,7 @@ def __init__(
self.logger.debug("logging_level set to %s", logging_level)

if base_url == "":
self.base_url = os.getenv("LLMWHISPERER_BASE_URL_V2", BASE_URL)
self.base_url = os.getenv("LLMWHISPERER_BASE_URL_V2", BASE_URL_V2)
else:
self.base_url = base_url
self.logger.debug("base_url set to %s", self.base_url)
Expand Down Expand Up @@ -281,7 +283,9 @@ def generate():
)
else:
params["url_in_post"] = True
req = requests.Request("POST", api_url, params=params, headers=self.headers, data=url)
req = requests.Request(
"POST", api_url, params=params, headers=self.headers, data=url
)
prepared = req.prepare()
s = requests.Session()
response = s.send(prepared, timeout=wait_timeout, stream=should_stream)
Expand All @@ -307,31 +311,41 @@ def generate():
message["extraction"] = {}
return message
if status["status"] == "processing":
self.logger.debug(f"Whisper-hash:{whisper_hash} | STATUS: processing...")
self.logger.debug(
f"Whisper-hash:{whisper_hash} | STATUS: processing..."
)
elif status["status"] == "delivered":
self.logger.debug(f"Whisper-hash:{whisper_hash} | STATUS: Already delivered!")
self.logger.debug(
f"Whisper-hash:{whisper_hash} | STATUS: Already delivered!"
)
raise LLMWhispererClientException(
{
"status_code": -1,
"message": "Whisper operation already delivered",
}
)
elif status["status"] == "unknown":
self.logger.debug(f"Whisper-hash:{whisper_hash} | STATUS: unknown...")
self.logger.debug(
f"Whisper-hash:{whisper_hash} | STATUS: unknown..."
)
raise LLMWhispererClientException(
{
"status_code": -1,
"message": "Whisper operation status unknown",
}
)
elif status["status"] == "failed":
self.logger.debug(f"Whisper-hash:{whisper_hash} | STATUS: failed...")
self.logger.debug(
f"Whisper-hash:{whisper_hash} | STATUS: failed..."
)
message["status_code"] = -1
message["message"] = "Whisper operation failed"
message["extraction"] = {}
return message
elif status["status"] == "processed":
self.logger.debug(f"Whisper-hash:{whisper_hash} | STATUS: processed!")
self.logger.debug(
f"Whisper-hash:{whisper_hash} | STATUS: processed!"
)
resultx = self.whisper_retrieve(whisper_hash=whisper_hash)
if resultx["status_code"] == 200:
message["status_code"] = 200
Expand Down

0 comments on commit 82baf42

Please sign in to comment.