From be91d4a9011010deb5756ea7890e2abf462ecd7a Mon Sep 17 00:00:00 2001
From: gdcsinaptik <giuseppe@sinaptik.ai>
Date: Fri, 13 Dec 2024 09:52:09 +0100
Subject: [PATCH] load and examples scripts

---
 new_examples/load_df.py    | 11 ++++++
 new_examples/quickstart.py | 24 +++++++++++++
 new_examples/save_csv.py   | 71 ++++++++++++++++++++++++++++++++++++++
 new_examples/use_openai.py | 27 +++++++++++++++
 pyproject.toml             |  1 +
 5 files changed, 134 insertions(+)
 create mode 100644 new_examples/load_df.py
 create mode 100644 new_examples/quickstart.py
 create mode 100644 new_examples/save_csv.py
 create mode 100644 new_examples/use_openai.py

diff --git a/new_examples/load_df.py b/new_examples/load_df.py
new file mode 100644
index 000000000..615a56440
--- /dev/null
+++ b/new_examples/load_df.py
@@ -0,0 +1,11 @@
+import os
+import pandasai as pai
+
+os.environ["PANDASAI_API_URL"] = "http://localhost:8000/"
+os.environ["PANDASAI_API_KEY"] = "PAI-test-key"
+
+# Load using organization/dataset format
+#df = pai.load("/home/giuseppe/Projects/pandas-ai/datasets/testing/loans")
+df = pai.load("/home/giuseppe/Projects/pandas-ai/datasets/testing/loans", virtualized=True)
+
+print(df.head())
\ No newline at end of file
diff --git a/new_examples/quickstart.py b/new_examples/quickstart.py
new file mode 100644
index 000000000..077c17101
--- /dev/null
+++ b/new_examples/quickstart.py
@@ -0,0 +1,24 @@
+import os
+import pandasai as pai
+
+os.environ["PANDASAI_API_URL"] = "http://localhost:8000/"
+os.environ["PANDASAI_API_KEY"] = "PAI-test-key"
+
+df = pai.read_csv("/home/giuseppe/Projects/pandas-ai/examples/data/Loan payments data.csv")
+
+# ask questions
+# replace "Which are the top 5 countries by sales?" with your question
+response =df.chat('How many loans are from men and have been paid off?')
+print(response)
+
+img =df.chat('Plot the chart of loans paid off by men vs by women')
+print(img)
+
+"""
+#minimal save
+df.save(
+    path="testing/loans",
+    name="loans",
+    description="Loans dataset"
+)
+"""
\ No newline at end of file
diff --git a/new_examples/save_csv.py b/new_examples/save_csv.py
new file mode 100644
index 000000000..48ad6fe9a
--- /dev/null
+++ b/new_examples/save_csv.py
@@ -0,0 +1,71 @@
+import os
+import pandasai as pai
+
+os.environ["PANDASAI_API_URL"] = "http://localhost:8000/"
+os.environ["PANDASAI_API_KEY"] = "PAI-test-key"
+
+df = pai.read_csv("/home/giuseppe/Projects/pandas-ai/examples/data/Loan payments data.csv")
+
+#save with fields descriptions
+df.save(
+    path="testing/loans",
+    name="loans",
+    description="Loans dataset",
+    columns=[
+        {
+            "name": "Loan_ID",
+            "type": "string",
+            "description": "Unique identifier for each loan"
+        },
+        {
+            "name": "loan_status",
+            "type": "string",
+            "description": "Status of the loan (PAIDOFF, COLLECTION, COLLECTION_PAIDOFF)"
+        },
+        {
+            "name": "Principal",
+            "type": "number",
+            "description": "The initial amount of the loan"
+        },
+        {
+            "name": "terms",
+            "type": "number",
+            "description": "The duration of the loan in days"
+        },
+        {
+            "name": "effective_date",
+            "type": "date",
+            "description": "The date when the loan became effective"
+        },
+        {
+            "name": "due_date",
+            "type": "date",
+            "description": "The date when the loan payment is due"
+        },
+        {
+            "name": "paid_off_time",
+            "type": "datetime",
+            "description": "The timestamp when the loan was paid off (if applicable)"
+        },
+        {
+            "name": "past_due_days",
+            "type": "number",
+            "description": "Number of days the payment is past due (if applicable)"
+        },
+        {
+            "name": "age",
+            "type": "number",
+            "description": "Age of the borrower"
+        },
+        {
+            "name": "education",
+            "type": "string",
+            "description": "Education level of the borrower (High School or Below, Bechalor, college, Master or Above)"
+        },
+        {
+            "name": "Gender",
+            "type": "string",
+            "description": "Gender of the borrower (male/female)"
+        }
+    ]
+)
\ No newline at end of file
diff --git a/new_examples/use_openai.py b/new_examples/use_openai.py
new file mode 100644
index 000000000..588d21e51
--- /dev/null
+++ b/new_examples/use_openai.py
@@ -0,0 +1,27 @@
+import os
+import pandasai as pai
+from pandasai_openai import OpenAI
+
+os.environ["PANDASAI_API_URL"] = "http://localhost:8000/"
+os.environ["PANDASAI_API_KEY"] = "PAI-test-key"
+
+llm = OpenAI(api_token="your-key")
+# Print LLM details
+print("LLM Type:", type(llm).__name__)
+print("LLM Instance:", llm)
+print("LLM Model:", llm.model)
+
+pai.config.set({
+   "llm": llm,
+})
+
+# Print current config to verify LLM setting
+print("\nCurrent PandasAI Config:")
+print("Active LLM:", pai.config._config.llm)
+
+df = pai.load("/home/giuseppe/Projects/pandas-ai/datasets/testing/loans")
+print(df.head())
+
+response = df.chat("What is the average age of the borrowers?")
+
+print(response)
\ No newline at end of file
diff --git a/pyproject.toml b/pyproject.toml
index c16abe69b..88f2364fe 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -11,6 +11,7 @@ packages = [{include = "pandasai"}]
 python = ">=3.8,<3.9.7 || >3.9.7,<3.12"
 python-dotenv = "^1.0.0"
 pandas = "^2.0.3"
+pyarrow = "^14.0.1"
 scipy = "1.10.1"
 astor = "^0.8.1"
 matplotlib = "^3.7.1"