phani-data-recon 1.0.7

Phani's Generic Data Reconciliation and Deduplication Utility

SAP ↔ Salesforce Data Reconciliation Utility

Reconcile SAP and Salesforce master data at bulk scale (300K–400K records). Produces a 10-tab Excel workbook and an HTML dashboard with KPIs, field-level diffs, fuzzy match candidates, and a prioritised action plan.

Supports multiple entity pairs through config, including:

• Accounts
• Orders
• Order Items
• Any custom SAP ↔ SF dataset with mapped keys and fields

Quick Start

Installation

# Install from PyPI
py -m pip install phani-data-recon

# Upgrade to the latest version
py -m pip install --upgrade phani-data-recon

# Verify installed version
py -m pip show phani-data-recon

Run the Application

Using the CLI command:

# Verify the CLI is available
reconcile-accounts --help

# Run with explicit SAP and Salesforce input files
reconcile-accounts --sap input/sap_accounts.csv --sf input/sf_accounts.csv

Using py module (alternative if CLI is not on PATH):

py -m phani_data_recon.cli --sap input/sap_accounts.csv --sf input/sf_accounts.csv

Configuration & Input/Output

Configure input and output folders via config file:

Edit config/rules.yaml to set default paths:

input:
  sap:
    directory: "input"
    file_name: "sap_accounts.csv"
  sf:
    directory: "input"
    file_name: "sf_accounts.csv"

output:
  formats: ["excel", "html"]
  report:
    directory: "output"
    file_name: "reconciliation_report"

Then run with config:

reconcile-accounts --config config/rules.yaml
py -m phani_data_recon.cli --config config/rules.yaml

Override output directory at runtime:

reconcile-accounts --config config/rules.yaml --output-dir output/custom_run
py -m phani_data_recon.cli --config config/rules.yaml --output-dir output/custom_run

Validation & Advanced Options

Validate headers and config only (dry-run):

reconcile-accounts --sap input/sap_accounts.csv --sf input/sf_accounts.csv --dry-run
py -m phani_data_recon.cli --sap input/sap_accounts.csv --sf input/sf_accounts.csv --dry-run

Generate only HTML output:

reconcile-accounts --sap input/sap_accounts.csv --sf input/sf_accounts.csv --formats html
py -m phani_data_recon.cli --sap input/sap_accounts.csv --sf input/sf_accounts.csv --formats html

Skip fuzzy matching (faster for large files):

reconcile-accounts --sap input/sap_accounts.csv --sf input/sf_accounts.csv --no-fuzzy
py -m phani_data_recon.cli --sap input/sap_accounts.csv --sf input/sf_accounts.csv --no-fuzzy

Enable verbose logging:

reconcile-accounts --sap input/sap_accounts.csv --sf input/sf_accounts.csv --verbose
py -m phani_data_recon.cli --sap input/sap_accounts.csv --sf input/sf_accounts.csv --verbose

Standalone Dedup Command (Separate from Reconciliation)

Use this command when you only want deduplication and do not want to run the full reconciliation flow. It works for Accounts, Orders, Order Items, and any other entity pair configured in YAML.

You can choose dedup logic per run:

• primary: dedup by primary key column
• fuzzy: dedup by similarity on one or more fields passed as CLI params

How fields are resolved for each mode:

• --dedup-mode primary uses --sap-primary-key / --sf-primary-key when passed; otherwise uses join.primary.sap_col / join.primary.sf_col from config.
• --dedup-mode fuzzy uses --sap-fuzzy-field / --sf-fuzzy-field when passed; otherwise uses defaults from config: fuzzy_match.fields[].sap_col for SAP and sf_fuzzy_dedup.fields[].col for SF.
• If fuzzy mode is selected and no fuzzy fields resolve for the selected entity, the run fails fast with a clear error.

# Accounts (default rules.yaml)
dedup-records --system sap --config config/rules.yaml --output-dir output/dedup
dedup-records --system sf  --config config/rules.yaml --output-dir output/dedup

# Orders
dedup-records --system sap --config config/rules.orders.example.yaml --output-dir output/dedup
dedup-records --system sf  --config config/rules.orders.example.yaml --output-dir output/dedup

# Order Items
dedup-records --system sap --config config/rules.order_items.example.yaml --output-dir output/dedup
dedup-records --system sf  --config config/rules.order_items.example.yaml --output-dir output/dedup

# Deduplicate only one side
dedup-records --system sap --sap input/sap_orders.csv --config config/rules.orders.example.yaml --output-dir output/dedup
dedup-records --system sf  --sf  input/sf_order_items.csv --config config/rules.order_items.example.yaml --output-dir output/dedup

# Fuzzy dedup with explicit entity fields (SF)
dedup-records --system sf --dedup-mode fuzzy --sf-fuzzy-field Name --sf-fuzzy-field WC_Email__c --fuzzy-min-score 85 --config config/rules.yaml --output-dir output/dedup

# Fuzzy dedup with explicit entity fields (SAP)
dedup-records --system sap --dedup-mode fuzzy --sap-fuzzy-field name1 --sap-fuzzy-field smtp_addr --fuzzy-min-score 85 --config config/rules.orders.example.yaml --output-dir output/dedup

# Fuzzy dedup with SAP fields passed from CLI
dedup-records --system sap --dedup-mode fuzzy --sap-fuzzy-field Order_Number --sap-fuzzy-field Customer_Name --fuzzy-min-score 85 --fuzzy-match-mode weighted --config config/rules.orders.example.yaml --output-dir output/dedup

# Fuzzy dedup with SF fields passed from CLI
dedup-records --system sf --dedup-mode fuzzy --sf-fuzzy-field OrderNumber --sf-fuzzy-field AccountName --fuzzy-min-score 85 --fuzzy-match-mode weighted --config config/rules.orders.example.yaml --output-dir output/dedup

# Module form (if console script is not on PATH)
py -m phani_data_recon.dedup_cli --system sf --sf input/sf_accounts.csv --config config/rules.yaml --output-dir output/dedup

Output files are generated as:

• output/dedup/<entity>_sap_deduped_<run_id>.csv
• output/dedup/<entity>_sap_duplicates_<run_id>.csv
• output/dedup/<entity>_sf_deduped_<run_id>.csv
• output/dedup/<entity>_sf_duplicates_<run_id>.csv

Where <entity> comes from entities.pair in config (slug format), for example:

• accounts_sap_deduped_<run_id>.csv
• orders_sf_duplicates_<run_id>.csv

Entity Command Cookbook

Use this section as a direct command reference for the two primary commands:

• Reconciliation command: reconcile-accounts
• Dedup-only command: dedup-records

Account Reconciliation + Dedup

# Reconciliation (Accounts)
reconcile-accounts --config config/rules.yaml

# Reconciliation with join-key overrides from CLI (also updates rules.yaml)
reconcile-accounts --config config/rules.yaml --sap-primary-key kunnr --sf-primary-key BP_PowerCerv_Account_Id__c --sap-fallback-key SAP_Unique_ID --sf-fallback-key WC_SAP_Identification__c --enable-fallback-key

# Dedup only (Accounts)
dedup-records --system sap --config config/rules.yaml --output-dir output/dedup/accounts
dedup-records --system sf  --config config/rules.yaml --output-dir output/dedup/accounts

# Dedup (Accounts) using fuzzy mode with explicit SF fields
dedup-records --system sf --dedup-mode fuzzy --sf-fuzzy-field Name --sf-fuzzy-field WC_Email__c --fuzzy-min-score 85 --fuzzy-match-mode any --config config/rules.yaml --output-dir output/dedup/accounts

Order Reconciliation + Dedup

# Reconciliation (Orders)
reconcile-accounts --config config/rules.orders.example.yaml

# Dedup only (Orders)
dedup-records --system sap --config config/rules.orders.example.yaml --output-dir output/dedup/orders
dedup-records --system sf  --config config/rules.orders.example.yaml --output-dir output/dedup/orders

# Dedup (Orders) using primary-key override
dedup-records --system sap --dedup-mode primary --sap-primary-key SAP_Order_Id --config config/rules.orders.example.yaml --output-dir output/dedup/orders
dedup-records --system sf  --dedup-mode primary --sf-primary-key External_Order_Id__c --config config/rules.orders.example.yaml --output-dir output/dedup/orders

# Dedup (Orders) using fuzzy fields passed per entity
dedup-records --system sap --dedup-mode fuzzy --sap-fuzzy-field Order_Number --sap-fuzzy-field Customer_Name --fuzzy-min-score 85 --config config/rules.orders.example.yaml --output-dir output/dedup/orders
dedup-records --system sf  --dedup-mode fuzzy --sf-fuzzy-field OrderNumber --sf-fuzzy-field AccountName --fuzzy-min-score 85 --config config/rules.orders.example.yaml --output-dir output/dedup/orders

Order Item Reconciliation + Dedup

# Reconciliation (Order Items)
reconcile-accounts --config config/rules.order_items.example.yaml

# Dedup only (Order Items)
dedup-records --system sap --config config/rules.order_items.example.yaml --output-dir output/dedup/order_items
dedup-records --system sf  --config config/rules.order_items.example.yaml --output-dir output/dedup/order_items

# Dedup (Order Items) using fuzzy fields passed per entity
dedup-records --system sap --dedup-mode fuzzy --sap-fuzzy-field Material --sap-fuzzy-field Item_Description --fuzzy-min-score 85 --config config/rules.order_items.example.yaml --output-dir output/dedup/order_items
dedup-records --system sf  --dedup-mode fuzzy --sf-fuzzy-field ProductCode --sf-fuzzy-field Description --fuzzy-min-score 85 --config config/rules.order_items.example.yaml --output-dir output/dedup/order_items

Module Form (No PATH dependency)

# Reconciliation command module form
py -m phani_data_recon.cli --config config/rules.orders.example.yaml

# Dedup command module form
py -m phani_data_recon.dedup_cli --system sf --config config/rules.order_items.example.yaml --output-dir output/dedup/order_items

Platform-specific path examples:

Windows:

reconcile-accounts --sap .\input\sap_accounts.csv --sf .\input\sf_accounts.csv
py -m phani_data_recon.cli --config .\config\rules.yaml --dry-run

# macOS
reconcile-accounts --sap ./input/sap_accounts.csv --sf ./input/sf_accounts.csv
python3 -m phani_data_recon.cli --config ./config/rules.yaml --dry-run

If Windows cmd does not recognize reconcile-accounts, add your Python Scripts directory to PATH and reopen cmd:

setx PATH "%PATH%;C:\Users\SeshaphaniBysani\AppData\Local\Python\pythoncore-3.14-64\Scripts"

Then verify:

where reconcile-accounts
reconcile-accounts --help

For local development in this repository, editable install still works:

py -m pip install -e .

Package Usage

# Explicit input files
reconcile-accounts --sap input/sap_accounts.csv --sf input/sf_accounts.csv

# Config-driven execution
reconcile-accounts --config config/rules.yaml

# Override only the output directory
reconcile-accounts --config config/rules.yaml --output-dir output/run_2026_05_11

# Generate only HTML output
reconcile-accounts --sap input/sap_accounts.csv --sf input/sf_accounts.csv --formats html

If the console script is not available on your PATH, use:

py -m phani_data_recon.cli --dry-run

Production verification on Windows cmd:

py -m pip show phani-data-recon
where reconcile-accounts
where dedup-records
py -m phani_data_recon.cli --sap input/sap_accounts.csv --sf input/sf_accounts.csv --dry-run

Expected state:

• Installed version should be 1.0.4 or later.
• If where reconcile-accounts is empty but module execution works, only PATH needs to be fixed.

Python API

Run reconciliation from another Python application:

from phani_data_recon.api import run_reconciliation

exit_code = run_reconciliation(
    sap="input/sap_accounts.csv",
    sf="input/sf_accounts.csv",
    config="config/rules.yaml",
    output_dir="output/api_run",
    formats=["excel", "html"],
    dry_run=False,
    no_fuzzy=False,
    verbose=True,
)

print(exit_code)

The API mirrors the CLI behavior and returns a process-style exit code.

Options

--sap         Path to SAP accounts CSV (optional if config input.sap is set)
--sf          Path to Salesforce accounts CSV (optional if config input.sf is set)
--config      Path to rules YAML (default: ./config/rules.yaml, then packaged default)
--output-dir  Output directory (default: from config)
--formats     excel html (default: both)
--sap-primary-key    Override SAP primary join key (join.primary.sap_col)
--sf-primary-key     Override SF primary join key (join.primary.sf_col)
--sap-fallback-key   Override SAP fallback key (join.fallback.sap_col)
--sf-fallback-key    Override SF fallback key (join.fallback.sf_col)
--enable-fallback-key   Enable fallback key matching
--disable-fallback-key  Disable fallback key matching
--dry-run     Validate config + headers only; no report written
--no-fuzzy    Skip fuzzy matching (faster for large files)
--verbose     Verbose logging

Path resolution precedence:

• If --sap / --sf are passed, CLI values are used.
• If not passed, values are resolved from config/rules.yaml under input.sap and input.sf.
• If join-key override args are passed (--sap-primary-key, --sf-primary-key, fallback key args), CLI values are used at runtime.
• When join-key override args are passed, rules.yaml is also updated with those keys.
• If --config is not passed, the CLI tries local config/rules.yaml first and then the packaged default config.
• If --output-dir is passed, it overrides output.report.directory.
• If neither CLI nor config provides paths, the run exits with an input path error.

Example key override run:

reconcile-accounts --config config/rules.yaml --sap-primary-key kunnr --sf-primary-key BP_PowerCerv_Account_Id__c --sap-fallback-key SAP_Unique_ID --sf-fallback-key WC_SAP_Identification__c --enable-fallback-key

Standalone dedup command options:

dedup-records --help

# Common options
--system sap|sf
--dedup-mode primary|fuzzy
--sap-primary-key <column>
--sf-primary-key <column>
--sap-fuzzy-field <column>   # repeat for multiple fields
--sf-fuzzy-field <column>    # repeat for multiple fields
--fuzzy-min-score <0-100>
--fuzzy-match-mode weighted|any

Dedup mode tips:

• Use primary for deterministic duplicate detection by business key.
• Use fuzzy when entity keys are unreliable and duplicate detection should be based on descriptive fields.
• For fuzzy, pass fields that are meaningful for that entity (for example: Account Name/Email for Accounts, Order Number/Account Name for Orders, Product/Description for Order Items).

Configuration

Edit config/rules.yaml to change:

• Default input files via input.sap and input.sf (directory + file_name)
• Join key columns (SAP ↔ SF linking fields)
• Fallback-key matching toggle via join.fallback.enabled (default: false = primary-key-only matching)
• Field comparison rules, severity levels, and normalize modes
• Deduplication strategy (keep_first / keep_last / flag_all)
• Fuzzy match threshold and fields
• Output formats and directory
• Output report location/name via output.report.directory + output.report.file_name

Generic Entity Configuration

This utility is generic. To reconcile other object types (for example SAP Orders ↔ Salesforce Orders or SAP Order Items ↔ Salesforce Order Items), update:

• entities labels (display text in reports)
• input.sap and input.sf file paths
• join.primary.sap_col and join.primary.sf_col
• field_mappings for the entity-specific columns
• fuzzy_match.fields for the entity-specific columns

Example entity labels:

entities:
	pair:      "Orders"
	sap:       "SAP Orders"
	sf:        "Salesforce Orders"
	sap_short: "SAP"
	sf_short:  "SF"

Ready-to-use example configs are included:

• config/rules.orders.example.yaml
• config/rules.order_items.example.yaml

Run examples:

# SAP Orders vs Salesforce Orders
reconcile-accounts --config config/rules.orders.example.yaml

# SAP Order Items vs Salesforce Order Items
reconcile-accounts --config config/rules.order_items.example.yaml

# Module form
py -m phani_data_recon.cli --config config/rules.orders.example.yaml

When using the package outside this repository, pass your own config file with --config if you do not want to rely on the packaged defaults.

Output Report Configuration

Use the output.report block in config/rules.yaml to control where reports are written and what base filename is used.

output:
	formats: ["excel", "html"]
	report:
		directory: "output/month_end"
		file_name: "customer_reconciliation"

This writes reports under output/month_end/ using customer_reconciliation as the base name, for example:

• output/month_end/customer_reconciliation_<run_id>.html
• output/month_end/customer_reconciliation_<run_id>.xlsx

Rules:

• --output-dir overrides output.report.directory
• output.report.file_name sets the report filename prefix
• output.formats selects Excel, HTML, or both

Example commands:

# Use output settings from config
reconcile-accounts --config config/rules.yaml

# Override only the output directory at runtime
reconcile-accounts --config config/rules.yaml --output-dir output/ad_hoc_run

Config Reference (Input + Join)

input:
	sap:
		directory: "input"
		file_name: "sap_accounts.csv"
	sf:
		directory: "input"
		file_name: "sf_accounts.csv"

join:
	primary:
		sap_col: "SAP_Unique_ID"
		sf_col:  "BP_PowerCerv_Account_Id__c"
	fallback:
		enabled: false
		sap_col: "SAP_Unique_ID"
		sf_col:  "WC_SAP_Identification__c"

output:
	formats: ["excel", "html"]
	report:
		directory: "output"
		file_name: "reconciliation_report"

Notes:

• Set join.fallback.enabled: false for strict primary-key-only matching (default).
• Set join.fallback.enabled: true only when you explicitly want fallback-key matching.

Report Tabs

Tab	Content
Summary	KPI counts, match rate, exception rate
Exact_Matches	Records found in both systems
Field_Mismatches	Field-level diffs (CRITICAL / HIGH / INFO)
SAP_Only	SAP records missing from Salesforce
SF_Only	Salesforce records missing from SAP
SAP_Duplicates	Duplicate SAP rows before dedup
SF_Duplicates	Duplicate SF rows before dedup
Fuzzy_Match_Candidates	Likely-same records not linked by ID
Data_Quality_Issues	Null IDs, bad formats, validation failures
Action_Plan	P1–P4 prioritised remediation table

Run Tests

py -m pip install pytest
py -m pytest tests/ -v

Distribution (Business Rollout)

# Build wheel + source distribution
py -m pip install build
py -m build

# Install locally from wheel
py -m pip install dist/phani_data_recon-1.0.4-py3-none-any.whl

If reconcile-accounts is not on PATH, run:

py -m phani_data_recon.cli --dry-run

Published package:

py -m pip install --upgrade phani-data-recon

Legacy script usage inside this repository still works:

python run_reconciliation.py --dry-run

CI Publishing (GitHub Actions)

This repository includes .github/workflows/publish-pypi.yml to publish new releases to PyPI without storing a PyPI API token in GitHub.

One-time PyPI setup:

• In PyPI, open the phani-data-recon project settings.
• Add a Trusted Publisher for this GitHub repository.
• Set the workflow name to publish-pypi.yml.
• Set the environment name to pypi.

Release flow:

• Bump the version in pyproject.toml.
• Create a GitHub release or run the workflow manually from the Actions tab.
• The workflow builds dist/ artifacts and publishes them with PyPI trusted publishing.

Notes:

• This workflow uses GitHub OIDC via id-token: write, so no TWINE_PASSWORD secret is required in GitHub.
• Keep local twine usage only for manual emergency releases.

Project Structure

reconciliation_project/
├── input/           ← Place source CSVs here
├── config/          ← rules.yaml + schema
├── src/             ← All Python modules
├── templates/       ← Jinja2 HTML template
├── tests/           ← pytest test suite
├── output/          ← Reports generated here
└── run_reconciliation.py