Merge branch 'main' of github.com:microsoft/qlib into fix_docs

.github/workflows/test_qlib_from_pip.yml

@@ -45,9 +45,6 @@ jobs:
 
     - name: Qlib installation test
       run: |
-        # 2024-05-30 scs has released a new version: 3.2.4.post2,
-        # This will cause the CI to fail, so we have limited the version of scs for now.
-        python -m pip install "scs<=3.2.4"
         python -m pip install pyqlib
 
     - name: Install Lightgbm for MacOS
@@ -68,8 +65,5 @@ jobs:
         cd qlib
 
     - name: Test workflow by config
-      # On macos-11 system, it will lead to "Segmentation fault: 11" error,
-      # which may be caused by the excessive memory overhead of macos-11 system, so we disable macos-11 temporarily here.
-      if: ${{ matrix.os != 'macos-11' }}
       run: |
         qrun examples/benchmarks/LightGBM/workflow_config_lightgbm_Alpha158.yaml

.github/workflows/test_qlib_from_source.yml

@@ -72,10 +72,8 @@ jobs:
         black . -l 120 --check --diff
 
     - name: Make html with sphinx
-      # Since read the docs builds on ubuntu 22.04, we only need to test that the build passes on ubuntu 22.04.
-      if: ${{ matrix.os == 'ubuntu-22.04' }}
       run: |
-        cd docs
+        cd docs 
         sphinx-build -W --keep-going -b html . _build
         cd ..
 
@@ -161,16 +159,11 @@ jobs:
 
     # Run after data downloads
     - name: Check Qlib ipynb with nbconvert
-      # Running the nbconvert check on a macos-11 system results in a "Kernel died" error, so we've temporarily disabled macos-11 here.
-      if: ${{ matrix.os != 'macos-11' }}
       run: |
         # add more ipynb files in future
         jupyter nbconvert --to notebook --execute examples/workflow_by_code.ipynb
 
     - name: Test workflow by config (install from source)
-      # On macos-11 system, it will lead to "Segmentation fault: 11" error,
-      # which may be caused by the excessive memory overhead of macos-11 system, so we disable macos-11 temporarily here.
-      if: ${{ matrix.os != 'macos-11' }}
       run: |
         python -m pip install numba
         python qlib/workflow/cli.py examples/benchmarks/LightGBM/workflow_config_lightgbm_Alpha158.yaml

README.md

@@ -40,7 +40,7 @@ Recent released features
 Features released before 2021 are not listed here.
 
 <p align="center">
-  <img src="docs/_static/img/logo/1.png" />
+  <img src="http://fintech.msra.cn/images_v070/logo/1.png" />
 </p>
 
 Qlib is an open-source, AI-oriented quantitative investment platform that aims to realize the potential, empower research, and create value using AI technologies in quantitative investment, from exploring ideas to implementing productions. Qlib supports diverse machine learning modeling paradigms, including supervised learning, market dynamics modeling, and reinforcement learning.
@@ -166,7 +166,7 @@ Also, users can install the latest dev version ``Qlib`` by the source code accor
 * Clone the repository and install ``Qlib`` as follows.
     ```bash
     git clone https://github.com/microsoft/qlib.git && cd qlib
-    pip install .  # `pip install -e .[dev]` is recommended for development. check details in docs/developer/code_standard_and_dev_guide.rst
+    pip install .
     ```
   **Note**:  You can install Qlib with `python setup.py install` as well. But it is not the recommended approach. It will skip `pip` and cause obscure problems. For example, **only** the command ``pip install .`` **can** overwrite the stable version installed by ``pip install pyqlib``, while the command ``python setup.py install`` **can't**.
 
@@ -175,20 +175,6 @@ Also, users can install the latest dev version ``Qlib`` by the source code accor
 **Tips for Mac**: If you are using Mac with M1, you might encounter issues in building the wheel for LightGBM, which is due to missing dependencies from OpenMP. To solve the problem, install openmp first with ``brew install libomp`` and then run ``pip install .`` to build it successfully. 
 
 ## Data Preparation
-❗ Due to more restrict data security policy. The offical dataset is disabled temporarily. You can try [this data source](https://github.com/chenditc/investment_data/releases) contributed by the community.
-Here is an example to download the data updated on 20220720.
-```bash
-wget https://github.com/chenditc/investment_data/releases/download/20220720/qlib_bin.tar.gz
-mkdir -p ~/.qlib/qlib_data/cn_data
-tar -zxvf qlib_bin.tar.gz -C ~/.qlib/qlib_data/cn_data --strip-components=2
-rm -f qlib_bin.tar.gz
-```
-
-The official dataset below will resume in short future.
-
-
-----
-
 Load and prepare data by running the following code:
 
 ### Get with module

docs/component/model.rst

@@ -86,7 +86,7 @@ Example
             },
         }
 
-        # model initialization
+        # model initiaiton
         model = init_instance_by_config(task["model"])
         dataset = init_instance_by_config(task["dataset"])
 

docs/requirements.txt

@@ -5,4 +5,3 @@ scipy
 scikit-learn
 pandas
 tianshou
-sphinx_rtd_theme

examples/orderbook_data/README.md

@@ -16,7 +16,7 @@ Current version of script with default value tries to connect localhost **via de
 
 Run following command to install necessary libraries
 ```
-pip install pytest coverage gdown
+pip install pytest coverage
 pip install arctic  # NOTE: pip may fail to resolve the right package dependency !!! Please make sure the dependency are satisfied.
 ```
 
@@ -27,8 +27,7 @@ pip install arctic  # NOTE: pip may fail to resolve the right package dependency
 2. Please follow following steps to download example data
 ```bash
 cd examples/orderbook_data/
-gdown https://drive.google.com/uc?id=15nZF7tFT_eKVZAcMFL1qPS4jGyJflH7e  # Proxies may be necessary here.
-python ../../scripts/get_data.py _unzip --file_path highfreq_orderbook_example_data.zip --target_dir .
+python ../../scripts/get_data.py download_data --target_dir . --file_name highfreq_orderbook_example_data.zip
 ```
 
 3. Please import the example data to your mongo db

examples/portfolio/README.md

@@ -20,7 +20,7 @@ We use China stock market data for our example.
 1. Prepare CSI300 weight:
 
    ```bash
-   wget https://github.com/SunsetWolf/qlib_dataset/releases/download/v0/csi300_weight.zip
+   wget http://fintech.msra.cn/stock_data/downloads/csi300_weight.zip
    unzip -d ~/.qlib/qlib_data/cn_data csi300_weight.zip
    rm -f csi300_weight.zip
    ```

examples/workflow_by_code.ipynb

@@ -161,7 +161,7 @@
     "    },\n",
     "}\n",
     "\n",
-    "# model initialization\n",
+    "# model initiaiton\n",
     "model = init_instance_by_config(task[\"model\"])\n",
     "dataset = init_instance_by_config(task[\"dataset\"])\n",
     "\n",

qlib/__init__.py

@@ -2,7 +2,7 @@
 # Licensed under the MIT License.
 from pathlib import Path
 
-__version__ = "0.9.5.99"
+__version__ = "0.9.4.99"
 __version__bak = __version__  # This version is backup for QlibConfig.reset_qlib_version
 import os
 from typing import Union

qlib/contrib/model/pytorch_alstm_ts.py

@@ -160,10 +160,6 @@ class ALSTM(Model):
 
         if self.metric in ("", "loss"):
             return -self.loss_fn(pred[mask], label[mask])
-        elif self.metric == "mse":
-            mask = ~torch.isnan(label)
-            weight = torch.ones_like(label)
-            return -self.mse(pred[mask], label[mask], weight[mask])
 
         raise ValueError("unknown metric `%s`" % self.metric)
 

qlib/data/client.py

@@ -35,7 +35,7 @@ class Client:
     def connect_server(self):
         """Connect to server."""
         try:
-            self.sio.connect(f"ws://{self.server_host}:{self.server_port}")
+            self.sio.connect("ws://" + self.server_host + ":" + str(self.server_port))
         except socketio.exceptions.ConnectionError:
             self.logger.error("Cannot connect to server - check your network or server status")
 

qlib/data/data.py

@@ -536,6 +536,7 @@ class DatasetProvider(abc.ABC):
         """
         if len(fields) == 0:
             raise ValueError("fields cannot be empty")
+        fields = fields.copy()
         column_names = [str(f) for f in fields]
         return column_names
 
@@ -616,7 +617,7 @@ class DatasetProvider(abc.ABC):
 
         data = pd.DataFrame(obj)
         if not data.empty and not np.issubdtype(data.index.dtype, np.dtype("M")):
-            # If the underlaying provides the data not in datetime format, we'll convert it into datetime format
+            # If the underlaying provides the data not in datatime formmat, we'll convert it into datetime format
             _calendar = Cal.calendar(freq=freq)
             data.index = _calendar[data.index.values.astype(int)]
         data.index.names = ["datetime"]

qlib/data/dataset/__init__.py

@@ -403,7 +403,7 @@ class TSDataSampler:
             np.full((1, self.data_arr.shape[1]), np.nan, dtype=self.data_arr.dtype),
             axis=0,
         )
-        self.nan_idx = len(self.data_arr) - 1  # The last line is all NaN; setting it to -1 can cause bug #1716
+        self.nan_idx = -1  # The last line is all NaN
 
         # the data type will be changed
         # The index of usable data is between start_idx and end_idx

qlib/data/dataset/utils.py

@@ -9,7 +9,7 @@ if TYPE_CHECKING:
     from qlib.data.dataset import DataHandler
 
 
-def get_level_index(df: pd.DataFrame, level: Union[str, int]) -> int:
+def get_level_index(df: pd.DataFrame, level=Union[str, int]) -> int:
     """
 
     get the level index of `df` given `level`

qlib/model/trainer.py

@@ -41,7 +41,7 @@ def _log_task_info(task_config: dict):
 
 def _exe_task(task_config: dict):
     rec = R.get_recorder()
-    # model & dataset initialization
+    # model & dataset initiation
     model: Model = init_instance_by_config(task_config["model"], accept_types=Model)
     dataset: Dataset = init_instance_by_config(task_config["dataset"], accept_types=Dataset)
     reweighter: Reweighter = task_config.get("reweighter", None)

qlib/tests/data.py

@@ -12,11 +12,15 @@ import datetime
 from tqdm import tqdm
 from pathlib import Path
 from loguru import logger
+from cryptography.fernet import Fernet
 from qlib.utils import exists_qlib_data
 
 
 class GetData:
-    REMOTE_URL = "https://github.com/SunsetWolf/qlib_dataset/releases/download"
+    REMOTE_URL = "https://qlibpublic.blob.core.windows.net/data/default/stock_data"
+    # "?" is not included in the token.
+    TOKEN = b"gAAAAABkmDhojHc0VSCDdNK1MqmRzNLeDFXe5hy8obHpa6SDQh4de6nW5gtzuD-fa6O_WZb0yyqYOL7ndOfJX_751W3xN5YB4-n-P22jK-t6ucoZqhT70KPD0Lf0_P328QPJVZ1gDnjIdjhi2YLOcP4BFTHLNYO0mvzszR8TKm9iT5AKRvuysWnpi8bbYwGU9zAcJK3x9EPL43hOGtxliFHcPNGMBoJW4g_ercdhi0-Qgv5_JLsV-29_MV-_AhuaYvJuN2dEywBy"
+    KEY = "EYcA8cgorA8X9OhyMwVfuFxn_1W3jGk6jCbs3L2oPoA="
 
     def __init__(self, delete_zip_file=False):
         """
@@ -29,45 +33,9 @@ class GetData:
         self.delete_zip_file = delete_zip_file
 
     def merge_remote_url(self, file_name: str):
-        """
-        Generate download links.
-
-        Parameters
-        ----------
-        file_name: str
-            The name of the file to be downloaded.
-            The file name can be accompanied by a version number, (e.g.: v2/qlib_data_simple_cn_1d_latest.zip),
-            if no version number is attached, it will be downloaded from v0 by default.
-        """
-        return f"{self.REMOTE_URL}/{file_name}" if "/" in file_name else f"{self.REMOTE_URL}/v0/{file_name}"
-
-    def download(self, url: str, target_path: [Path, str]):
-        """
-        Download a file from the specified url.
-
-        Parameters
-        ----------
-        url: str
-            The url of the data.
-        target_path: str
-            The location where the data is saved, including the file name.
-        """
-        file_name = str(target_path).rsplit("/", maxsplit=1)[-1]
-        resp = requests.get(url, stream=True, timeout=60)
-        resp.raise_for_status()
-        if resp.status_code != 200:
-            raise requests.exceptions.HTTPError()
-
-        chunk_size = 1024
-        logger.warning(
-            f"The data for the example is collected from Yahoo Finance. Please be aware that the quality of the data might not be perfect. (You can refer to the original data source: https://finance.yahoo.com/lookup.)"
-        )
-        logger.info(f"{os.path.basename(file_name)} downloading......")
-        with tqdm(total=int(resp.headers.get("Content-Length", 0))) as p_bar:
-            with target_path.open("wb") as fp:
-                for chunk in resp.iter_content(chunk_size=chunk_size):
-                    fp.write(chunk)
-                    p_bar.update(chunk_size)
+        fernet = Fernet(self.KEY)
+        token = fernet.decrypt(self.TOKEN).decode()
+        return f"{self.REMOTE_URL}/{file_name}?{token}"
 
     def download_data(self, file_name: str, target_dir: [Path, str], delete_old: bool = True):
         """
@@ -102,7 +70,21 @@ class GetData:
         target_path = target_dir.joinpath(_target_file_name)
 
         url = self.merge_remote_url(file_name)
-        self.download(url=url, target_path=target_path)
+        resp = requests.get(url, stream=True, timeout=60)
+        resp.raise_for_status()
+        if resp.status_code != 200:
+            raise requests.exceptions.HTTPError()
+
+        chunk_size = 1024
+        logger.warning(
+            f"The data for the example is collected from Yahoo Finance. Please be aware that the quality of the data might not be perfect. (You can refer to the original data source: https://finance.yahoo.com/lookup.)"
+        )
+        logger.info(f"{os.path.basename(file_name)} downloading......")
+        with tqdm(total=int(resp.headers.get("Content-Length", 0))) as p_bar:
+            with target_path.open("wb") as fp:
+                for chunk in resp.iter_content(chunk_size=chunk_size):
+                    fp.write(chunk)
+                    p_bar.update(chunk_size)
 
         self._unzip(target_path, target_dir, delete_old)
         if self.delete_zip_file:
@@ -117,9 +99,7 @@ class GetData:
         return status
 
     @staticmethod
-    def _unzip(file_path: [Path, str], target_dir: [Path, str], delete_old: bool = True):
-        file_path = Path(file_path)
-        target_dir = Path(target_dir)
+    def _unzip(file_path: Path, target_dir: Path, delete_old: bool = True):
         if delete_old:
             logger.warning(
                 f"will delete the old qlib data directory(features, instruments, calendars, features_cache, dataset_cache): {target_dir}"

qlib/workflow/task/utils.py

@@ -242,7 +242,7 @@ class TimeAdjuster:
 
     def shift(self, seg: tuple, step: int, rtype=SHIFT_SD) -> tuple:
         """
-        Shift the datetime of segment
+        Shift the datatime of segment
 
         If there are None (which indicates unbounded index) in the segment, this method will return None.
 

scripts/data_collector/base.py

@@ -301,7 +301,6 @@ class Normalize:
             na_values={col: symbol_na if col == self._symbol_field_name else default_na for col in columns},
         )
 
-        # NOTE: It has been reported that there may be some problems here, and the specific issues will be dealt with when they are identified.
         df = self._normalize_obj.normalize(df)
         if df is not None and not df.empty:
             if self._end_date is not None:

scripts/data_collector/cn_index/requirements.txt

@@ -5,5 +5,3 @@ pandas
 lxml
 loguru
 tqdm
-yahooquery
-openpyxl

scripts/data_collector/crypto/README.md

@@ -9,7 +9,7 @@ pip install -r requirements.txt
 ```
 
 ## Usage of the dataset
-> *Crypto dataset only support Data retrieval function but not support backtest function due to the lack of OHLC data.*
+> *Crypto dateset only support Data retrieval function but not support backtest function due to the lack of OHLC data.*
 
 ## Collector Data
 

scripts/data_collector/utils.py

@@ -15,6 +15,7 @@ from typing import Iterable, Tuple, List
 
 import numpy as np
 import pandas as pd
+from lxml import etree
 from loguru import logger
 from yahooquery import Ticker
 from tqdm import tqdm
@@ -189,43 +190,17 @@ def get_hs_stock_symbols() -> list:
     global _HS_SYMBOLS  # pylint: disable=W0603
 
     def _get_symbol():
-        """
-        Get the stock pool from a web page and process it into the format required by yahooquery.
-        Format of data retrieved from the web page: 600519, 000001
-        The data format required by yahooquery: 600519.ss, 000001.sz
-
-        Returns
-        -------
-            set: Returns the set of symbol codes.
-
-        Examples:
-        -------
-            {600000.ss, 600001.ss, 600002.ss, 600003.ss, ...}
-        """
-        url = "http://99.push2.eastmoney.com/api/qt/clist/get?pn=1&pz=10000&po=1&np=1&fs=m:0+t:6,m:0+t:80,m:1+t:2,m:1+t:23,m:0+t:81+s:2048&fields=f12"
-        try:
-            resp = requests.get(url, timeout=None)
-            resp.raise_for_status()
-        except requests.exceptions.HTTPError as e:
-            raise requests.exceptions.HTTPError(f"Request to {url} failed with status code {resp.status_code}") from e
-
-        try:
-            _symbols = [_v["f12"] for _v in resp.json()["data"]["diff"]]
-        except Exception as e:
-            logger.warning("An error occurred while extracting data from the response.")
-            raise
-
-        if len(_symbols) < 3900:
-            raise ValueError("The complete list of stocks is not available.")
-
-        # Add suffix after the stock code to conform to yahooquery standard, otherwise the data will not be fetched.
-        _symbols = [
-            _symbol + ".ss" if _symbol.startswith("6") else _symbol + ".sz" if _symbol.startswith(("0", "3")) else None
-            for _symbol in _symbols
-        ]
-        _symbols = [_symbol for _symbol in _symbols if _symbol is not None]
-
-        return set(_symbols)
+        _res = set()
+        for _k, _v in (("ha", "ss"), ("sa", "sz"), ("gem", "sz")):
+            resp = requests.get(HS_SYMBOLS_URL.format(s_type=_k), timeout=None)
+            _res |= set(
+                map(
+                    lambda x: "{}.{}".format(re.findall(r"\d+", x)[0], _v),  # pylint: disable=W0640
+                    etree.HTML(resp.text).xpath("//div[@class='result']/ul//li/a/text()"),  # pylint: disable=I1101
+                )
+            )
+            time.sleep(3)
+        return _res
 
     if _HS_SYMBOLS is None:
         symbols = set()

scripts/data_collector/yahoo/collector.py

@@ -796,9 +796,6 @@ class Run(BaseRun):
             # get 1m data
             $ python collector.py download_data --source_dir ~/.qlib/stock_data/source --region CN --start 2020-11-01 --end 2020-11-10 --delay 0.1 --interval 1m
         """
-        if self.interval == "1d" and pd.Timestamp(end) > pd.Timestamp(datetime.datetime.now().strftime("%Y-%m-%d")):
-            raise ValueError(f"end_date: {end} is greater than the current date.")
-
         super(Run, self).download_data(max_collector_count, delay, start, end, check_data_length, limit_nums)
 
     def normalize_data(

setup.py

@@ -46,7 +46,7 @@ if not _CYTHON_INSTALLED:
 REQUIRED = [
     "numpy>=1.12.0, <1.24",
     "pandas>=0.25.1",
-    "scipy>=1.7.3",
+    "scipy>=1.0.0",
     "requests>=2.18.0",
     "sacred>=0.7.4",
     "python-socketio",
@@ -82,7 +82,7 @@ REQUIRED = [
     "dill",
     "dataclasses;python_version<'3.7'",
     "filelock",
-    "jinja2",
+    "jinja2<3.1.0",  # for passing the readthedocs workflow.
     "gym",
     # Installing the latest version of protobuf for python versions below 3.8 will cause unit tests to fail.
     "protobuf<=3.20.1;python_version<='3.8'",
@@ -166,9 +166,6 @@ setup(
             "lxml",
             "baostock",
             "yahooquery",
-            # 2024-05-30 scs has released a new version: 3.2.4.post2,
-            # this version, causes qlib installation to fail, so we've limited the scs version a bit for now.
-            "scs<=3.2.4",
             "beautifulsoup4",
             # In version 0.4.11 of tianshou, the code:
             # logits, hidden = self.actor(batch.obs, state=state, info=batch.info)

tests/data_mid_layer_tests/test_dataset.py

@@ -5,9 +5,8 @@ import unittest
 import pytest
 import sys
 from qlib.tests import TestAutoData
-from qlib.data.dataset import TSDatasetH, TSDataSampler
+from qlib.data.dataset import TSDatasetH
 import numpy as np
-import pandas as pd
 import time
 from qlib.data.dataset.handler import DataHandlerLP
 
@@ -99,54 +98,6 @@ class TestDataset(TestAutoData):
             print(idx[i])
 
 
-class TestTSDataSampler(unittest.TestCase):
-    def test_TSDataSampler(self):
-        """
-        Test TSDataSampler for issue #1716
-        """
-        datetime_list = ["2000-01-31", "2000-02-29", "2000-03-31", "2000-04-30", "2000-05-31"]
-        instruments = ["000001", "000002", "000003", "000004", "000005"]
-        index = pd.MultiIndex.from_product(
-            [pd.to_datetime(datetime_list), instruments], names=["datetime", "instrument"]
-        )
-        data = np.random.randn(len(datetime_list) * len(instruments))
-        test_df = pd.DataFrame(data=data, index=index, columns=["factor"])
-        dataset = TSDataSampler(test_df, datetime_list[0], datetime_list[-1], step_len=2)
-        print()
-        print("--------------dataset[0]--------------")
-        print(dataset[0])
-        print("--------------dataset[1]--------------")
-        print(dataset[1])
-        assert len(dataset[0]) == 2
-        self.assertTrue(np.isnan(dataset[0][0]))
-        self.assertEqual(dataset[0][1], dataset[1][0])
-        self.assertEqual(dataset[1][1], dataset[2][0])
-        self.assertEqual(dataset[2][1], dataset[3][0])
-
-    def test_TSDataSampler2(self):
-        """
-        Extra test TSDataSampler to prevent incorrect filling of nan for the values at the front
-        """
-        datetime_list = ["2000-01-31", "2000-02-29", "2000-03-31", "2000-04-30", "2000-05-31"]
-        instruments = ["000001", "000002", "000003", "000004", "000005"]
-        index = pd.MultiIndex.from_product(
-            [pd.to_datetime(datetime_list), instruments], names=["datetime", "instrument"]
-        )
-        data = np.random.randn(len(datetime_list) * len(instruments))
-        test_df = pd.DataFrame(data=data, index=index, columns=["factor"])
-        dataset = TSDataSampler(test_df, datetime_list[2], datetime_list[-1], step_len=3)
-        print()
-        print("--------------dataset[0]--------------")
-        print(dataset[0])
-        print("--------------dataset[1]--------------")
-        print(dataset[1])
-        for i in range(3):
-            self.assertFalse(np.isnan(dataset[0][i]))
-            self.assertFalse(np.isnan(dataset[1][i]))
-        self.assertEqual(dataset[0][1], dataset[1][0])
-        self.assertEqual(dataset[0][2], dataset[1][1])
-
-
 if __name__ == "__main__":
     unittest.main(verbosity=10)
 

tests/test_all_pipeline.py

@@ -27,7 +27,7 @@ def train(uri_path: str = None):
             model performance
     """
 
-    # model initialization
+    # model initiaiton
     model = init_instance_by_config(CSI300_GBDT_TASK["model"])
     dataset = init_instance_by_config(CSI300_GBDT_TASK["dataset"])
     # To test __repr__