inspecty 0.1.0__tar.gz

inspecty-0.1.0/PKG-INFO

@@ -0,0 +1,137 @@
+Metadata-Version: 2.1
+Name: inspecty
+Version: 0.1.0
+Summary: Universal data inspector — one function to understand any dataset
+Home-page: https://github.com/manjur-ai/inspecty
+Author: Manjur Alam
+Classifier: Programming Language :: Python :: 3
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Operating System :: OS Independent
+Classifier: Topic :: Scientific/Engineering
+Classifier: Topic :: Software Development :: Libraries :: Python Modules
+Requires-Python: >=3.9
+Description-Content-Type: text/markdown
+
+# inspecty
+
+**Universal data inspector — one function to understand any dataset.**
+
+```bash
+pip install inspecty
+```
+
+Instead of `print()`, `type()`, `.info()`, `.describe()` — just run `ins.inspect(data)` and get a high-density statistical report for any data type.
+
+---
+
+## Supported data types
+
+DataFrame, Series, NumPy array (1D/2D/3D), list, tuple, set, dictionary, JSON string, CSV file path, Excel file path — any Python object.
+
+## Quick example
+
+```python
+import inspecty as ins
+import pandas as pd
+
+df = pd.DataFrame({
+    "price":  [100.5, 101.2, 100.8, 102.1, 101.5],
+    "volume": [1000, 1500, 1200, 1800, 2000],
+})
+
+ins.inspect(df)
+```
+
+Output:
+
+```
+Data type: <class 'pandas.DataFrame'>
+Total rows: 5. Total columns: 2. Summary:
+
+column:      price      volume
+ dtype:    float64       int64
+     -      -----       -----
+ Row 0:      100.5        1000
+ Row 1:      101.2        1500
+ Row 2:      100.8        1200
+----------  -----       -----
+ count:          5           5
+   max:      102.1      2000.0
+   min:      100.5      1000.0
+  mean:     101.22      1500.0
+   std:   0.622093  412.310563
+median:      101.2      1500.0
+  mode:      100.5      1000.0
+   nan:          0           0
+```
+
+---
+
+## More examples
+
+### Pandas Series
+
+```python
+series = pd.Series([10, 20, 30, 40, 50], name="RSI")
+ins.inspect(series)
+```
+
+### Dictionary pivot
+
+```python
+data = {
+    "Ticker": ["BTC", "ETH", "SOL"],
+    "Price":  [62000.5, 3400.2, 145.1],
+    "Signal": ["Buy", "Hold", "Buy"],
+}
+ins.inspect(data)
+```
+
+### 3D NumPy tensor
+
+```python
+tensor = np.random.randn(2, 100, 5)  # (Tickers, Days, Features)
+ins.inspect(tensor)
+```
+
+### Auto-load files
+
+```python
+ins.inspect("historical_prices.csv")
+ins.inspect("data.xlsx")
+```
+
+### JSON string
+
+```python
+json_resp = '{"Symbol": ["AAPL", "TSLA"], "Price": [150.2, 700.5]}'
+ins.inspect(json_resp)
+```
+
+### System info
+
+```python
+ins.inspect("all_info")
+```
+
+---
+
+## Parameters
+
+| Param | Type | Default | Description |
+|-------|------|---------|-------------|
+| `df` | Any | required | Data object, file path, or JSON string |
+| `count` | int | 3 | Rows shown from top & bottom |
+| `silent` | bool | False | Hide headers, show only table |
+
+---
+
+## Why not `df.describe()`?
+
+`df.describe()` works only on numeric columns, hides NaN counts, and needs multiple calls for basic stats. `inspect()` handles mixed types, shows NaN explicitly, and works on any data structure — not just DataFrames.
+
+---
+
+## License
+
+MIT

inspecty-0.1.0/README.md

@@ -0,0 +1,123 @@
+# inspecty
+
+**Universal data inspector — one function to understand any dataset.**
+
+```bash
+pip install inspecty
+```
+
+Instead of `print()`, `type()`, `.info()`, `.describe()` — just run `ins.inspect(data)` and get a high-density statistical report for any data type.
+
+---
+
+## Supported data types
+
+DataFrame, Series, NumPy array (1D/2D/3D), list, tuple, set, dictionary, JSON string, CSV file path, Excel file path — any Python object.
+
+## Quick example
+
+```python
+import inspecty as ins
+import pandas as pd
+
+df = pd.DataFrame({
+    "price":  [100.5, 101.2, 100.8, 102.1, 101.5],
+    "volume": [1000, 1500, 1200, 1800, 2000],
+})
+
+ins.inspect(df)
+```
+
+Output:
+
+```
+Data type: <class 'pandas.DataFrame'>
+Total rows: 5. Total columns: 2. Summary:
+
+column:      price      volume
+ dtype:    float64       int64
+     -      -----       -----
+ Row 0:      100.5        1000
+ Row 1:      101.2        1500
+ Row 2:      100.8        1200
+----------  -----       -----
+ count:          5           5
+   max:      102.1      2000.0
+   min:      100.5      1000.0
+  mean:     101.22      1500.0
+   std:   0.622093  412.310563
+median:      101.2      1500.0
+  mode:      100.5      1000.0
+   nan:          0           0
+```
+
+---
+
+## More examples
+
+### Pandas Series
+
+```python
+series = pd.Series([10, 20, 30, 40, 50], name="RSI")
+ins.inspect(series)
+```
+
+### Dictionary pivot
+
+```python
+data = {
+    "Ticker": ["BTC", "ETH", "SOL"],
+    "Price":  [62000.5, 3400.2, 145.1],
+    "Signal": ["Buy", "Hold", "Buy"],
+}
+ins.inspect(data)
+```
+
+### 3D NumPy tensor
+
+```python
+tensor = np.random.randn(2, 100, 5)  # (Tickers, Days, Features)
+ins.inspect(tensor)
+```
+
+### Auto-load files
+
+```python
+ins.inspect("historical_prices.csv")
+ins.inspect("data.xlsx")
+```
+
+### JSON string
+
+```python
+json_resp = '{"Symbol": ["AAPL", "TSLA"], "Price": [150.2, 700.5]}'
+ins.inspect(json_resp)
+```
+
+### System info
+
+```python
+ins.inspect("all_info")
+```
+
+---
+
+## Parameters
+
+| Param | Type | Default | Description |
+|-------|------|---------|-------------|
+| `df` | Any | required | Data object, file path, or JSON string |
+| `count` | int | 3 | Rows shown from top & bottom |
+| `silent` | bool | False | Hide headers, show only table |
+
+---
+
+## Why not `df.describe()`?
+
+`df.describe()` works only on numeric columns, hides NaN counts, and needs multiple calls for basic stats. `inspect()` handles mixed types, shows NaN explicitly, and works on any data structure — not just DataFrames.
+
+---
+
+## License
+
+MIT

inspecty-0.1.0/inspecty/__init__.py

@@ -0,0 +1,4 @@
+from .printdata import inspect
+
+__all__ = ["inspect"]
+__version__ = "0.1.0"

inspecty-0.1.0/inspecty/printdata.py

@@ -0,0 +1,310 @@
+import numpy as np
+import pandas as pd
+import os
+import json
+import sys
+import subprocess
+import platform
+import shutil
+import importlib.metadata
+from pandas.api.types import is_numeric_dtype
+
+# ============================================================
+# ENVIRONMENT INSPECTION (Silent Mode Enabled)
+# ============================================================
+def _inspect_all_info():
+    # Setup for silent subprocess execution on Windows (prevents CMD flicker)
+    startupinfo = None
+    if platform.system() == "Windows":
+        startupinfo = subprocess.STARTUPINFO()
+        startupinfo.dwFlags |= subprocess.STARTF_USESHOWWINDOW
+        startupinfo.wShowWindow = subprocess.SW_HIDE
+
+    print("\n========== SYSTEM ==========")
+    print("OS:          ", platform.system(), platform.release())
+    print("Architecture:", platform.machine())
+    print("Processor:   ", platform.processor())
+
+    # --- LIGHTWEIGHT GPU & CUDA DETECTION ---
+    print("\n========== GPU & CUDA ==========")
+    try:
+        gpu_raw = subprocess.check_output(
+            ["nvidia-smi", "--query-gpu=gpu_name,memory.total,driver_version", "--format=csv,noheader,nounits"],
+            text=True, 
+            startupinfo=startupinfo
+        ).strip()
+        
+        if gpu_raw:
+            name, mem, driver = gpu_raw.split(',')
+            print(f"Hardware:       {name.strip()}")
+            print(f"Total Memory:   {mem.strip()} MB")
+            print(f"Driver Version: {driver.strip()}")
+            
+            cuda_info = subprocess.check_output(["nvidia-smi"], text=True, startupinfo=startupinfo)
+            for line in cuda_info.split('\n'):
+                if "CUDA Version:" in line:
+                    v_part = line.split("CUDA Version:")[1].strip()
+                    print(f"CUDA Version:   {v_part.split(' ')[0]}")
+                    break
+    except:
+        print("Status:         No NVIDIA GPU or Driver detected via nvidia-smi.")
+
+    print("\n========== SCRIPT ==========")
+    print("Current Working Directory:", os.getcwd())
+    try:
+        print("Script Path:              ", os.path.abspath(sys.argv[0]))
+    except: pass
+
+    # --- UPDATED ENVIRONMENT LOGIC (ml_env detection) ---
+    print("\n========== CURRENT ENVIRONMENT ==========")
+    is_venv = (hasattr(sys, 'real_prefix') or 
+               (getattr(sys, 'base_prefix', sys.prefix) != sys.prefix))
+    
+    if is_venv:
+        env_name = os.path.basename(sys.prefix)
+    else:
+        env_name = "Global System"
+
+    print("Name:             ", env_name)
+    print("Path:             ", sys.prefix)
+    print("Python:           ", sys.version.split()[0])
+    print("Python Executable:", sys.executable)
+
+    try:
+        pip_info = subprocess.check_output(
+            [sys.executable, "-m", "pip", "--version"],
+            text=True, startupinfo=startupinfo
+        ).strip()
+        print("\nPip:              ", pip_info)
+    except:
+        print("\nPip:               Not detected")
+
+    print("\nInstalled Packages:")
+    try:
+        packages = sorted(
+            [(d.metadata["Name"], d.version) for d in importlib.metadata.distributions()],
+            key=lambda x: x[0].lower()
+        )
+        line = []
+        for name, ver in packages:
+            line.append(f"{name} {ver}")
+            if len(line) == 4:
+                print(" | ".join(line))
+                line = []
+        if line:
+            print(" | ".join(line))
+    except:
+        print("Could not read installed packages")
+
+    print("\n========== DEFAULT PYTHON (CMD) ==========")
+    try:
+        default_python = shutil.which("python")
+        if default_python:
+            print("python command ->", default_python)
+            try:
+                v = subprocess.check_output(
+                    [default_python, "--version"],
+                    text=True, startupinfo=startupinfo
+                ).strip()
+                print("Version:         ", v)
+            except: pass
+        else:
+            print("python command not found in PATH")
+    except: pass
+
+    print("\n========== OTHER PYTHON INSTALLATIONS ==========")
+    versions = set()
+    try:
+        result = subprocess.run(
+            ["where", "python"],
+            capture_output=True, text=True, startupinfo=startupinfo
+        )
+        python_paths = list(set(result.stdout.splitlines()))
+        current_exe = os.path.abspath(sys.executable)
+        idx = 1
+
+        for path in python_paths:
+            if os.path.abspath(path) == current_exe:
+                continue
+
+            print(f"\nEnvironment #{idx}")
+            print("Python Path:", path)
+            try:
+                v = subprocess.check_output(
+                    [path, "--version"],
+                    text=True, startupinfo=startupinfo
+                ).strip()
+                print("Python:     ", v)
+                versions.add(v)
+            except: pass
+
+            try:
+                pipv = subprocess.check_output(
+                    [path, "-m", "pip", "--version"],
+                    text=True, startupinfo=startupinfo
+                ).strip()
+                print("pip:        ", pipv)
+            except:
+                print("pip:         not installed")
+
+            # RESTORED PACKAGE LISTING FOR OTHER ENVIRONMENTS
+            print("\nInstalled Packages:")
+            try:
+                pkg_output = subprocess.check_output(
+                    [path, "-m", "pip", "list", "--format=freeze"],
+                    text=True, startupinfo=startupinfo
+                )
+                pkg_lines = sorted(pkg_output.strip().splitlines())
+                line = []
+                for p in pkg_lines:
+                    name_ver = p.replace("==", " ")
+                    line.append(name_ver)
+                    if len(line) == 4:
+                        print(" | ".join(line))
+                        line = []
+                if line:
+                    print(" | ".join(line))
+            except:
+                print("Could not read installed packages")
+            
+            idx += 1
+    except:
+        print("Could not detect other python installations")
+
+    print("\n========== WARNINGS ==========")
+    if len(versions) > 1:
+        print("WARNING: multiple Python versions detected")
+        for v in versions:
+            print(" ", v)
+
+    print("\n========== TOOL ==========")
+    tool = "Terminal"
+    if "VSCODE_PID" in os.environ: tool = "VS Code"
+    elif "PYCHARM_HOSTED" in os.environ: tool = "PyCharm"
+    elif "ipykernel" in sys.modules: tool = "Jupyter"
+    elif "idlelib" in sys.modules: tool = "IDLE"
+    print("Running in:", tool)
+
+# ============================================================
+# ORIGINAL INSPECT FUNCTION
+# ============================================================
+def inspect(df, count=None, silent=False):
+    if isinstance(df, str) and df.lower() == "all_info":
+        _inspect_all_info()
+        return
+
+    if not silent:
+        print(f"\nData type: {type(df)}")
+
+    is_native_type = False
+    if isinstance(df, str):
+        if df.lower().endswith(('.csv', '.xlsx')):
+            if os.path.exists(df):
+                if not silent: print(f"File detected: '{df}'. Loading for inspection...")
+                try:
+                    df = pd.read_csv(df) if df.lower().endswith('.csv') else pd.read_excel(df)
+                except Exception as e:
+                    if not silent: print(f"Error loading file: {e}")
+            else:
+                if not silent: print(f"Path string detected, but file not found at: {df}")
+        elif (df.strip().startswith('{') and df.strip().endswith('}')) or \
+             (df.strip().startswith('[') and df.strip().endswith(']')):
+            try:
+                df = json.loads(df)
+                if not silent: print("JSON string detected. Parsed for inspection.")
+            except: pass
+
+    if isinstance(df, dict):
+        is_native_type = True
+        lengths = [len(v) for v in df.values() if isinstance(v, (list, tuple))]
+        all_lists = all(isinstance(v, (list, tuple)) for v in df.values())
+        if all_lists and len(set(lengths)) == 1 and len(lengths) > 0:
+            df = pd.DataFrame(df)
+            if not silent: print("Dictionary detected with equal-length lists. Dictionary keys used as Columns.")
+        else:
+            df = pd.Series(df).to_frame(name="value")
+            if not silent: print("Input is a dictionary. Keys moved to Index for inspection.")
+    elif isinstance(df, (list, tuple, set)):
+        is_native_type = True
+        original_type_name = type(df).__name__ 
+        df = pd.Series(list(df))
+        if not silent: print(f"Input is a {original_type_name}. Converted for inspection.")
+
+    if isinstance(df, (pd.DataFrame, pd.Series, np.ndarray)):   
+        if isinstance(df, pd.Series): 
+            series_name = df.name if df.name else "it_is_series"
+            new_df = df.to_frame(name=series_name)
+            if not silent:
+                prefix = "" if is_native_type else "It is a pandas Series. "
+                print(f"{prefix}Total items: {len(new_df)}. Summary:")
+        elif isinstance(df, np.ndarray): 
+            numpy_new = df.copy()
+            if numpy_new.ndim == 1:        
+                new_df = pd.DataFrame(numpy_new, columns=["numpy_array"])
+                if not silent: print(f"Input is a 1D numpy array (Shape: {numpy_new.shape}). Total items: {len(new_df)}. Summary:")
+            elif numpy_new.ndim == 2:
+                col_names = [f"col_{i}" for i in range(numpy_new.shape[1])]
+                new_df = pd.DataFrame(numpy_new, columns=col_names)
+                if not silent: print(f"Input is a 2D numpy array (Shape: {numpy_new.shape}). Total rows: {len(new_df)}. Total columns: {len(new_df.columns)}. Summary:")
+            elif numpy_new.ndim == 3:
+                depth, rows, cols = numpy_new.shape
+                new_df = pd.DataFrame(numpy_new.reshape(-1, cols), columns=[f"feature_{i}" for i in range(cols)])
+                if not silent: print(f"Input is a 3D numpy array (Shape: {numpy_new.shape}). Total items (flattened): {depth*rows}. Total columns: {cols}. Summary:")
+            else:
+                if not silent: print(f"Array is {numpy_new.ndim}D (Shape: {numpy_new.shape}). Default print:")
+                print(numpy_new); return
+        elif isinstance(df, pd.DataFrame):
+            new_df = df.copy()
+            if not silent:
+                prefix = "" if is_native_type else "It is a pandas DataFrame. "
+                print(f"{prefix}Total rows: {len(new_df)}. Total columns: {len(new_df.columns)}. Summary:")
+
+        def safe_stat(col, func_name):
+            try:
+                if func_name == 'max':
+                    val = col.max()
+                    return val.strftime('%Y-%m-%d %H:%M:%S') if pd.api.types.is_datetime64_any_dtype(col) else val
+                if func_name == 'min':
+                    val = col.min()
+                    return val.strftime('%Y-%m-%d %H:%M:%S') if pd.api.types.is_datetime64_any_dtype(col) else val
+                if func_name == 'mean': return col.mean() if is_numeric_dtype(col) else "-"
+                if func_name == 'std': return col.std() if is_numeric_dtype(col) else "-"
+                if func_name == 'median': return col.median() if is_numeric_dtype(col) else "-"
+                if func_name == 'mode':
+                    m = col.mode()
+                    return m.iloc[0] if not m.empty else "-"
+            except: return "-"
+            return "-"
+
+        row_to_print = count if count else 3
+        if new_df.empty: print("The DataFrame is empty."); return
+        if not is_numeric_dtype(new_df.index) or (len(new_df) > 0 and new_df.index[-1] != (len(new_df) - 1)):
+            new_df.reset_index(drop=False, inplace=True)
+
+        headers = new_df.columns
+        dashs = ['-----'] * len(headers)
+        summary_rows = [["dtype:", *new_df.dtypes.astype(str)], ["-", *dashs]]
+        
+        first_rows = new_df.head(min(row_to_print, len(new_df)))
+        for i, (_, row) in enumerate(first_rows.iterrows()):
+            summary_rows.append([f"Row {i}:", *row.tolist()])
+        
+        if len(new_df) > (row_to_print * 2):
+            summary_rows.append(["--", *dashs])
+            last_rows = new_df.tail(row_to_print)
+            for i, (_, row) in enumerate(last_rows.iterrows()):
+                summary_rows.append([f"Row {len(new_df)-row_to_print+i}:", *row.tolist()])
+                
+        summary_rows.append(["----------", *dashs])
+        summary_rows.append(["count:", *new_df.count().tolist()])
+        summary_rows.append(["max:", *new_df.apply(lambda c: safe_stat(c, 'max')).tolist()])
+        summary_rows.append(["min:", *new_df.apply(lambda c: safe_stat(c, 'min')).tolist()])
+        summary_rows.append(["mean:", *new_df.apply(lambda c: safe_stat(c, 'mean')).tolist()])
+        summary_rows.append(["std:", *new_df.apply(lambda c: safe_stat(c, 'std')).tolist()])
+        summary_rows.append(["median:", *new_df.apply(lambda c: safe_stat(c, 'median')).tolist()])
+        summary_rows.append(["mode:", *new_df.apply(lambda c: safe_stat(c, 'mode')).tolist()])
+        summary_rows.append(["nan:", *new_df.isna().sum().tolist()])
+        
+        print(pd.DataFrame(summary_rows, columns=["column:", *headers]).to_string(index=False))
+    else:
+        print(df)

inspecty-0.1.0/inspecty.egg-info/PKG-INFO

@@ -0,0 +1,137 @@
+Metadata-Version: 2.1
+Name: inspecty
+Version: 0.1.0
+Summary: Universal data inspector — one function to understand any dataset
+Home-page: https://github.com/manjur-ai/inspecty
+Author: Manjur Alam
+Classifier: Programming Language :: Python :: 3
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Operating System :: OS Independent
+Classifier: Topic :: Scientific/Engineering
+Classifier: Topic :: Software Development :: Libraries :: Python Modules
+Requires-Python: >=3.9
+Description-Content-Type: text/markdown
+
+# inspecty
+
+**Universal data inspector — one function to understand any dataset.**
+
+```bash
+pip install inspecty
+```
+
+Instead of `print()`, `type()`, `.info()`, `.describe()` — just run `ins.inspect(data)` and get a high-density statistical report for any data type.
+
+---
+
+## Supported data types
+
+DataFrame, Series, NumPy array (1D/2D/3D), list, tuple, set, dictionary, JSON string, CSV file path, Excel file path — any Python object.
+
+## Quick example
+
+```python
+import inspecty as ins
+import pandas as pd
+
+df = pd.DataFrame({
+    "price":  [100.5, 101.2, 100.8, 102.1, 101.5],
+    "volume": [1000, 1500, 1200, 1800, 2000],
+})
+
+ins.inspect(df)
+```
+
+Output:
+
+```
+Data type: <class 'pandas.DataFrame'>
+Total rows: 5. Total columns: 2. Summary:
+
+column:      price      volume
+ dtype:    float64       int64
+     -      -----       -----
+ Row 0:      100.5        1000
+ Row 1:      101.2        1500
+ Row 2:      100.8        1200
+----------  -----       -----
+ count:          5           5
+   max:      102.1      2000.0
+   min:      100.5      1000.0
+  mean:     101.22      1500.0
+   std:   0.622093  412.310563
+median:      101.2      1500.0
+  mode:      100.5      1000.0
+   nan:          0           0
+```
+
+---
+
+## More examples
+
+### Pandas Series
+
+```python
+series = pd.Series([10, 20, 30, 40, 50], name="RSI")
+ins.inspect(series)
+```
+
+### Dictionary pivot
+
+```python
+data = {
+    "Ticker": ["BTC", "ETH", "SOL"],
+    "Price":  [62000.5, 3400.2, 145.1],
+    "Signal": ["Buy", "Hold", "Buy"],
+}
+ins.inspect(data)
+```
+
+### 3D NumPy tensor
+
+```python
+tensor = np.random.randn(2, 100, 5)  # (Tickers, Days, Features)
+ins.inspect(tensor)
+```
+
+### Auto-load files
+
+```python
+ins.inspect("historical_prices.csv")
+ins.inspect("data.xlsx")
+```
+
+### JSON string
+
+```python
+json_resp = '{"Symbol": ["AAPL", "TSLA"], "Price": [150.2, 700.5]}'
+ins.inspect(json_resp)
+```
+
+### System info
+
+```python
+ins.inspect("all_info")
+```
+
+---
+
+## Parameters
+
+| Param | Type | Default | Description |
+|-------|------|---------|-------------|
+| `df` | Any | required | Data object, file path, or JSON string |
+| `count` | int | 3 | Rows shown from top & bottom |
+| `silent` | bool | False | Hide headers, show only table |
+
+---
+
+## Why not `df.describe()`?
+
+`df.describe()` works only on numeric columns, hides NaN counts, and needs multiple calls for basic stats. `inspect()` handles mixed types, shows NaN explicitly, and works on any data structure — not just DataFrames.
+
+---
+
+## License
+
+MIT

inspecty-0.1.0/inspecty.egg-info/SOURCES.txt

@@ -0,0 +1,9 @@
+README.md
+setup.py
+inspecty/__init__.py
+inspecty/printdata.py
+inspecty.egg-info/PKG-INFO
+inspecty.egg-info/SOURCES.txt
+inspecty.egg-info/dependency_links.txt
+inspecty.egg-info/requires.txt
+inspecty.egg-info/top_level.txt

inspecty-0.1.0/inspecty.egg-info/dependency_links.txt

@@ -0,0 +1 @@
+

inspecty-0.1.0/inspecty.egg-info/requires.txt

@@ -0,0 +1,3 @@
+numpy>=1.21.0
+pandas>=1.3.0
+openpyxl>=3.0.0

inspecty-0.1.0/inspecty.egg-info/top_level.txt

@@ -0,0 +1 @@
+inspecty

inspecty-0.1.0/setup.cfg

@@ -0,0 +1,4 @@
+[egg_info]
+tag_build = 
+tag_date = 0
+

inspecty-0.1.0/setup.py

@@ -0,0 +1,28 @@
+from setuptools import setup, find_packages
+
+with open("README.md", "r", encoding="utf-8") as fh:
+    long_description = fh.read()
+
+setup(
+    name="inspecty",
+    version="0.1.0",
+    author="Manjur Alam",
+    description="Universal data inspector — one function to understand any dataset",
+    long_description=long_description,
+    long_description_content_type="text/markdown",
+    url="https://github.com/manjur-ai/inspecty",
+    packages=find_packages(),
+    install_requires=[
+        "numpy>=1.21.0",
+        "pandas>=1.3.0",
+        "openpyxl>=3.0.0",
+    ],
+    classifiers=[
+        "Programming Language :: Python :: 3",
+        "License :: OSI Approved :: MIT License",
+        "Operating System :: OS Independent",
+        "Topic :: Scientific/Engineering",
+        "Topic :: Software Development :: Libraries :: Python Modules",
+    ],
+    python_requires='>=3.9',
+)