Azure / Azure/azure-sdk-for-python

Batch analysis result not retrievable

オープン
#44,572 コメント 6 件 リアクション 1 件 担当者 0 名 GitHub で見る
customer-reported Document Intelligence needs-team-attention question Service Attention
主要言語
Python
スター
5.6k
フォーク
3.4k
平均マージ
2日
マージ済み PR(30日)
217

説明

- **Package Name**: azure-ai-documentintelligence
- **Package Version**: 1.0.2
- **Operating System**: Windows 10
- **Python Version**: 3.13.11

**Describe the bug**
I tried following the sample tutorials [sample_analyze_batch_documents.py](https://github.com/Azure/azure-sdk-for-python/blob/d47bfafcc4f72d08911f87f0ab34b57f5cdc72f4/sdk/documentintelligence/azure-ai-documentintelligence/samples/sample_analyze_batch_documents.py#L48) and [sample_analyze_batch_documents_async.py](https://github.com/Azure/azure-sdk-for-python/blob/d47bfafcc4f72d08911f87f0ab34b57f5cdc72f4/sdk/documentintelligence/azure-ai-documentintelligence/samples/async_samples/sample_analyze_batch_documents_async.py#L49), and in both cases, the mechanism to verify whether the operation is concluded does not work.
Specifically, I always get the message "The batch analyze is still in process..." even when I know for a fact that the analysis has concluded.

I also tried implementing a solution by using the method [add_done_callback](https://learn.microsoft.com/en-us/python/api/azure-core/azure.core.polling.lropoller?view=azure-python#azure-core-polling-lropoller-add-done-callback) but this also does not work. Unfortunately, there is no sample that illustrates the correct use of this method to monitor a batch analysis.

**To Reproduce**
Steps to reproduce the behavior:
1. run the samples.

**Expected behavior**
The prints should reflect the status of the batch analysis operation.

コントリビューションガイド

コントリビューションガイドを開く

調査の方向性

Start by running sample_analyze_batch_documents.py and async_samples/sample_analyze_batch_documents_async.py in the azure-ai-documentintelligence package, focusing on how completion status is checked. Read the LROPoller add_done_callback entry point mentioned in the report and compare its behavior with the samples. Done means the samples and callback path accurately reflect when batch analysis has completed.

索引モデルが issue の本文から書いたものです。

評価

技術スタック
azure, python
領域
api
issue の種類
バグ
難易度
3/5
見積もり時間
1〜2日
活発さ
静か
明瞭さ
おおむね明確
初心者へのやさしさ
55/100

新しい issue をメールで受け取る

初心者向けの GitHub issue を短くまとめたダイジェスト。