python -m venv .venv
source .venv/bin/activate
pip install --upgrade --quiet -r requirements.txt
https://python.langchain.com/docs/integrations/llms/bedrock/
Claude v3 models are not supported by this LLM.Please use `from langchain_community.chat_models import BedrockChat` instead.
BedrockはBedrockChatに変更
https://python.langchain.com/docs/get_started/quickstart/
この辺りも参考
https://python.langchain.com/docs/use_cases/sql/quickstart/
ModuleNotFoundError: No module named 'MySQLdb'
https://github.com/PyMySQL/mysqlclient/blob/main/README.md#macos-homebrew
以前、mysql-clientをmac portsでインストールした気がするので、割愛
ここで、mysqlclientをrequirements.txtに追加して、下記実行
pip install --upgrade --quiet -r requirements.txt
Exception: Can not find valid pkg-config name.
みたいなエラーが出てうまくいかない
sudo port install pkgconfig
package名と違うのやだね。
which pkg-config
/opt/local/bin/pkg-config
export PKG_CONFIG_PATH="/opt/local/lib/mysql8/pkgconfig"
うーんうまく帰ってこない
To get the total number of products, we can run the following SQL query:
SELECT COUNT(*) AS total_products
FROM products;
langchainが裏で生成してくれているプロンプトを確認
You are a MySQL expert. Given an input question, first create a syntactically correct MySQL query to run, then look at the results of the query and return the answer to the input question.
Unless the user specifies in the question a specific number of examples to obtain, query for at most 5 results using the LIMIT clause as per MySQL. You can order the results to return the most informative data in the database.
Never query for all columns from a table. You must query only the columns that are needed to answer the question. Wrap each column name in backticks (`) to denote them as delimited identifiers.
Pay attention to use only the column names you can see in the tables below. Be careful to not query for columns that do not exist. Also, pay attention to which column is in which table.
Pay attention to use CURDATE() function to get the current date, if the question involves "today".
Use the following format:
Question: Question here
SQLQuery: SQL Query to run
SQLResult: Result of the SQLQuery
Answer: Final answer here
Only use the following tables:
{table_info}
Question: {input}
LangChainDeprecationWarning: The class `BedrockChat` was deprecated in LangChain 0.0.34 and will be removed in 0.3. An updated version of the class exists in the langchain-aws package and should be used instead. To use it run `pip install -U langchain-aws` and import as `from langchain_aws import ChatBedrock`.
LangChainDeprecationWarning: The method `Chain.__call__` was deprecated in langchain 0.1.0 and will be removed in 0.2.0. Use invoke instead.
poetry run which python
を実行すると.venv/bin/python
を指し示す
poetry add $( cat requirements.txt )
でやってみたけど、長くなりすぎてダメな気がする
alembic init alembic
alembic upgrade head
alembic upgrade head --sql > migration.sql
alembic revision -m "create product table"
alembic revision --autogenerate -m "add price on products table"
- モデルを作る
- omymodelsで生成してコミットしてみる
- models.pyは別で作成したものをコピペ
- 上記の2つを比べてみる
(.venv) langchain$ poetry add omymodels
Using version ^0.17.0 for omymodels
Updating dependencies
Resolving dependencies... (0.1s)
Because no versions of omymodels match >0.17.0,<0.18.0
and omymodels (0.17.0) depends on pydantic (>=1.8.2,<2.0.0), omymodels (>=0.17.0,<0.18.0) requires pydantic (>=1.8.2,<2.0.0).
So, because langchain-samples depends on both pydantic (2.7.0) and omymodels (^0.17.0), version solving failed.
pyproject.tomlから削除
pydantic = "2.7.0"
pydantic-core = "2.18.1"
今度は成功
$ poetry add omymodels
Using version ^0.17.0 for omymodels
Updating dependencies
Resolving dependencies... (1.3s)
Package operations: 8 installs, 1 update, 1 removal
- Removing pydantic-core (2.18.1)
- Downgrading pydantic (2.7.0 -> 1.10.15)
- Installing regex (2024.5.10)
- Installing parsimonious (0.10.0)
- Installing ply (3.11)
- Installing jinja2 (3.1.4)
- Installing py-models-parser (0.7.0)
- Installing simple-ddl-parser (1.3.0)
- Installing table-meta (0.1.5)
- Installing omymodels (0.17.0)
ddlをdumpして、omymodelsで生成
ddlはalembic upgrade head --sql > migration.sql
で出てくるけど、
別で追加したテーブルが反映されてないので、
alembic revision --autogenerate -m "create helo table"
をする必要があると。
なんやかんやあって、データは投入できた