banksonsale-data 0.1.0__tar.gz

banksonsale_data-0.1.0/CITATION.cff

@@ -0,0 +1,17 @@
+cff-version: 1.2.0
+message: "If you use BanksOnSale Data, cite this software and the FDIC source and reporting dates recorded in the dataset manifest."
+title: "BanksOnSale Data"
+type: software
+authors:
+  - name: "BanksOnSale"
+version: 0.1.0
+date-released: 2026-10-29
+license: MIT
+url: "https://www.banksonsale.com/data"
+repository-code: "https://github.com/anshuman-pandey/banksonsale-data"
+abstract: "Keyless U.S. bank financial panels with explicit calendar quarters, metric units, archived FDIC responses and portable source-linked exports."
+keywords:
+  - bank data
+  - FDIC
+  - bank research
+  - reproducible research

banksonsale_data-0.1.0/DATA_LICENSE.md

@@ -0,0 +1,24 @@
+# BanksOnSale Data source and reuse
+
+Original package code is licensed under MIT. BanksOnSale's original dataset
+compilation and metadata in the pinned sample resources are offered under
+[Creative Commons Attribution 4.0](https://creativecommons.org/licenses/by/4.0/).
+FDIC BankFind Suite is the original regulatory data source. BanksOnSale does not
+claim exclusive ownership of the underlying government observations or imply
+FDIC endorsement. Consult the original source documentation for its terms.
+
+Recommended credit: **BanksOnSale Data, sourced from FDIC BankFind Suite**, followed
+by the bank CERT, reporting dates, package version and dataset ID. Raw source
+responses, request URLs, source-index creation timestamps and retrieval timestamps
+are retained in each sample. The bank's original reporting date is a different
+thing from the source index or retrieval date.
+
+These licenses do not cover unrelated market feeds, private BanksOnSale accounts,
+saved research, subscription-only data or third-party material elsewhere on
+BanksOnSale.com. This release contains only selected public FDIC observations and
+original code/metadata. The source and the compilation may contain errors;
+independently verify material conclusions against the filings.
+
+Source definitions: https://api.fdic.gov/banks/docs/risview_properties.yaml
+API documentation: https://api.fdic.gov/banks/docs/
+Corrections: https://www.banksonsale.com/corrections

banksonsale_data-0.1.0/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 BanksOnSale
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

banksonsale_data-0.1.0/MANIFEST.in

@@ -0,0 +1,7 @@
+include LICENSE DATA_LICENSE.md CITATION.cff README.md
+recursive-include src/banksonsale_data/resources *.json
+recursive-include docs *.md *.json
+recursive-include notebooks *.ipynb
+recursive-include examples *.csv *.json
+recursive-include tests *.py
+recursive-include scripts *.py

banksonsale_data-0.1.0/PKG-INFO

@@ -0,0 +1,190 @@
+Metadata-Version: 2.4
+Name: banksonsale-data
+Version: 0.1.0
+Summary: BanksOnSale Data: source-linked U.S. bank financial panels for research, classes, and applications.
+Author: BanksOnSale
+License-Expression: MIT AND CC-BY-4.0
+Project-URL: Homepage, https://banksonsale.com
+Project-URL: Repository, https://github.com/anshuman-pandey/banksonsale-data
+Project-URL: Documentation, https://github.com/anshuman-pandey/banksonsale-data#readme
+Project-URL: Issues, https://github.com/anshuman-pandey/banksonsale-data/issues
+Keywords: banksonsale,bank-data,fdic,banking,financial-data
+Classifier: Development Status :: 3 - Alpha
+Classifier: Intended Audience :: Developers
+Classifier: Intended Audience :: Education
+Classifier: Intended Audience :: Financial and Insurance Industry
+Classifier: Programming Language :: Python :: 3
+Classifier: Topic :: Office/Business :: Financial
+Requires-Python: >=3.10
+Description-Content-Type: text/markdown
+License-File: LICENSE
+License-File: DATA_LICENSE.md
+Provides-Extra: pandas
+Requires-Dist: pandas>=1.5; extra == "pandas"
+Provides-Extra: parquet
+Requires-Dist: pyarrow>=14; extra == "parquet"
+Provides-Extra: polars
+Requires-Dist: polars>=0.20; extra == "polars"
+Provides-Extra: io
+Requires-Dist: pandas>=1.5; extra == "io"
+Requires-Dist: pyarrow>=14; extra == "io"
+Provides-Extra: notebooks
+Requires-Dist: pandas>=1.5; extra == "notebooks"
+Requires-Dist: nbformat>=5; extra == "notebooks"
+Requires-Dist: nbclient>=0.10; extra == "notebooks"
+Requires-Dist: ipykernel>=6; extra == "notebooks"
+Dynamic: license-file
+
+# BanksOnSale Data
+
+**U.S. bank financial data with the quarter, units and original source attached.**
+
+[BanksOnSale.com](https://www.banksonsale.com) · [Public data catalog](https://www.banksonsale.com/data) · [Metric definitions](https://github.com/anshuman-pandey/banksonsale-data/blob/v0.1.0/docs/METRICS.md) · [OpenBB extension](https://github.com/anshuman-pandey/banksonsale-data/tree/v0.1.0/integrations/openbb)
+
+Get a bank's quarterly financial panel, inspect the source evidence, and export
+CSV, JSON or Parquet. The core package uses Python's standard library: **no
+BanksOnSale account, API key, paid subscription or background telemetry**.
+
+Start offline with real, pinned samples. Use the same interface for a live FDIC
+request when you want to choose another institution or reporting window.
+
+## Install v0.1.0
+
+Python 3.10 or newer. Install the versioned release wheel without requiring Git:
+
+```bash
+python -m pip install https://github.com/anshuman-pandey/banksonsale-data/releases/download/v0.1.0/banksonsale_data-0.1.0-py3-none-any.whl
+```
+
+Or install the tagged source with Git:
+
+```bash
+python -m pip install "banksonsale-data @ git+https://github.com/anshuman-pandey/banksonsale-data.git@v0.1.0"
+```
+
+These instructions use the GitHub release; they do not assume a PyPI release exists.
+Optional extras include `pandas`, `polars`, `parquet`, and `notebooks`.
+For example, install `pandas` in the same environment to use `.to_pandas()`.
+
+## Useful output in four lines
+
+```python
+from banksonsale_data import load_sample
+
+panel = load_sample(58665)  # Live Oak Banking Company; eight quarters through 2026Q2
+print(panel.rows[-1])
+panel.export("live-oak.csv")  # also writes live-oak.manifest.json with sources and units
+```
+
+The offline sample covers **2024-09-30 through 2026-06-30**. It is a frozen
+example, not a claim about the latest available filing. Three sample institutions
+are included: CERT 628, 3510 and 58665. They demonstrate the interface across
+bank sizes; they are not a matched peer group or a ranking.
+
+## Choose a bank and an exact reporting window
+
+```python
+from banksonsale_data import bank, search
+
+matches = search("Live Oak", state="NC")  # inspect matches and select a CERT
+panel = bank(58665).financials(quarters=8, through="2026Q2")
+print(panel.document["bank"]["name"], panel.document["coverage"])
+print(panel.dataset_id)
+```
+
+Omitting `through` asks FDIC for that bank's latest available reporting date and
+records that default in the manifest. Different banks can have different latest
+dates. Select the same explicit quarter when comparing banks. Requests are
+bounded and transient failures have capped retries.
+
+## Use your existing analysis tools
+
+```python
+frame = panel.to_pandas()       # optional pandas; evidence retained in frame.attrs["banksonsale"]
+records = panel.rows            # independent copy of standard Python records
+panel.export("live-oak.json")  # self-contained snapshot with archived evidence
+panel.export("live-oak.parquet")  # optional pyarrow; evidence in schema metadata
+```
+
+CSV works in spreadsheets and DuckDB. `.to_polars()` is available with Polars
+installed. See [the data contract](https://github.com/anshuman-pandey/banksonsale-data/blob/v0.1.0/docs/DATA_CONTRACT.md) for nulls, units and
+snapshot integrity. JSON from `panel.export()` is a BankPanel envelope; the
+website's quarter-wide JSON download is a separate documented format.
+
+## Run a real example
+
+- [Bank financials and exports](https://github.com/anshuman-pandey/banksonsale-data/blob/v0.1.0/notebooks/01_bank_financials.ipynb) — [Open in Colab](https://colab.research.google.com/github/anshuman-pandey/banksonsale-data/blob/v0.1.0/notebooks/01_bank_financials.ipynb).
+- [Same-quarter comparison and SQL](https://github.com/anshuman-pandey/banksonsale-data/blob/v0.1.0/notebooks/02_compare_banks.ipynb) — [Open in Colab](https://colab.research.google.com/github/anshuman-pandey/banksonsale-data/blob/v0.1.0/notebooks/02_compare_banks.ipynb).
+- [Ready-to-open comparison CSV](https://github.com/anshuman-pandey/banksonsale-data/blob/v0.1.0/examples/comparison-2026-q2.csv), with the [source manifest](https://github.com/anshuman-pandey/banksonsale-data/blob/v0.1.0/examples/comparison-2026-q2.manifest.json).
+
+The notebooks run locally without a cloud account. Their install cell retrieves
+the tagged public release. Colab has its own sign-in requirements and limits.
+
+## What you get
+
+Fourteen reported fields cover assets, total and domestic deposits, equity,
+intangibles, net loans, estimated uninsured deposits, core deposits, CET1,
+leverage, annualized ROA/ROE, domestic offices and employees.
+
+- Monetary columns end in `_usd` and are **U.S. dollars**, normalized once from FDIC's USD-thousands values.
+- Ratios end in `_pct`: **12.5 means 12.5%**. ROA and ROE are already annualized by the source.
+- CET1 uses **RBCT1CER**. Leverage uses **RBC1AAJ**. They remain distinct.
+- Missing observations stay `None`. Reported zero and negative observations remain unchanged.
+- Missing quarters remain explicit rows with `report_available=False`; older reports are never substituted.
+- FDIC CERT identifies an institution, not a listed holding company, ticker or individual branch.
+
+This is a small, versioned release of selected FDIC fields. It does not cover all
+bank data, customers' accounts, live transactions, full Call Reports, transaction
+prices or verified sale listings. Names and institution metadata reflect the
+upstream source at retrieval, not a reconstructed historical identity timeline.
+
+## Command line
+
+```bash
+banksonsale-data sample --cert 58665 --output live-oak.csv
+banksonsale-data search "Live Oak" --state NC
+banksonsale-data financials --cert 58665 --quarters 8 --through 2026Q2 --output live-oak.json
+banksonsale-data definitions
+```
+
+## Continue the research in BanksOnSale
+
+[Open Live Oak's public financial summary](https://www.banksonsale.com/banks/58665)
+or [explore the BanksOnSale research workspace](https://www.banksonsale.com/discover).
+The hosted product adds saved screens, watchlists, research dossiers, monitoring,
+models and exports; some features require a subscription. Its data edition is
+displayed separately and can differ from this package's selected period.
+
+These are optional links. The package keeps working without an account, sends
+no usage telemetry, and does not create a subscription. See [continuing your
+research](https://github.com/anshuman-pandey/banksonsale-data/blob/v0.1.0/docs/CONTINUE_IN_BANKSONSALE.md).
+
+## OpenBB
+
+The optional [independently maintained OpenBB V5 extension](https://github.com/anshuman-pandey/banksonsale-data/tree/v0.1.0/integrations/openbb)
+adds `obb.banksonsale`. It reuses this package and its definitions. It is not
+bundled with or endorsed by OpenBB. The core package does not require OpenBB.
+
+## Source, license and citation
+
+Original source: [FDIC BankFind Suite](https://api.fdic.gov/banks/docs/).
+BanksOnSale is an independent compiler; no FDIC affiliation or endorsement is implied.
+
+Original code: [MIT](https://github.com/anshuman-pandey/banksonsale-data/blob/v0.1.0/LICENSE). Sample compilation and original metadata:
+[CC BY 4.0, with separate source attribution](https://github.com/anshuman-pandey/banksonsale-data/blob/v0.1.0/DATA_LICENSE.md).
+Use [CITATION.cff](https://github.com/anshuman-pandey/banksonsale-data/blob/v0.1.0/CITATION.cff) and record the bank CERT, report dates and dataset ID.
+SHA-256 checks detect changes to a snapshot; they are not a government signature
+or proof that a filing is economically correct.
+
+## Contributing and verification
+
+See [CONTRIBUTING.md](https://github.com/anshuman-pandey/banksonsale-data/blob/v0.1.0/CONTRIBUTING.md). Offline checks:
+
+```bash
+python -m pip install -e .
+python -m unittest discover -s tests -v
+```
+
+Network refreshes are explicit maintainer actions, not part of the default tests.
+[Report a package issue](https://github.com/anshuman-pandey/banksonsale-data/issues)
+or [a source-data correction](https://www.banksonsale.com/corrections).

banksonsale_data-0.1.0/README.md

@@ -0,0 +1,153 @@
+# BanksOnSale Data
+
+**U.S. bank financial data with the quarter, units and original source attached.**
+
+[BanksOnSale.com](https://www.banksonsale.com) · [Public data catalog](https://www.banksonsale.com/data) · [Metric definitions](https://github.com/anshuman-pandey/banksonsale-data/blob/v0.1.0/docs/METRICS.md) · [OpenBB extension](https://github.com/anshuman-pandey/banksonsale-data/tree/v0.1.0/integrations/openbb)
+
+Get a bank's quarterly financial panel, inspect the source evidence, and export
+CSV, JSON or Parquet. The core package uses Python's standard library: **no
+BanksOnSale account, API key, paid subscription or background telemetry**.
+
+Start offline with real, pinned samples. Use the same interface for a live FDIC
+request when you want to choose another institution or reporting window.
+
+## Install v0.1.0
+
+Python 3.10 or newer. Install the versioned release wheel without requiring Git:
+
+```bash
+python -m pip install https://github.com/anshuman-pandey/banksonsale-data/releases/download/v0.1.0/banksonsale_data-0.1.0-py3-none-any.whl
+```
+
+Or install the tagged source with Git:
+
+```bash
+python -m pip install "banksonsale-data @ git+https://github.com/anshuman-pandey/banksonsale-data.git@v0.1.0"
+```
+
+These instructions use the GitHub release; they do not assume a PyPI release exists.
+Optional extras include `pandas`, `polars`, `parquet`, and `notebooks`.
+For example, install `pandas` in the same environment to use `.to_pandas()`.
+
+## Useful output in four lines
+
+```python
+from banksonsale_data import load_sample
+
+panel = load_sample(58665)  # Live Oak Banking Company; eight quarters through 2026Q2
+print(panel.rows[-1])
+panel.export("live-oak.csv")  # also writes live-oak.manifest.json with sources and units
+```
+
+The offline sample covers **2024-09-30 through 2026-06-30**. It is a frozen
+example, not a claim about the latest available filing. Three sample institutions
+are included: CERT 628, 3510 and 58665. They demonstrate the interface across
+bank sizes; they are not a matched peer group or a ranking.
+
+## Choose a bank and an exact reporting window
+
+```python
+from banksonsale_data import bank, search
+
+matches = search("Live Oak", state="NC")  # inspect matches and select a CERT
+panel = bank(58665).financials(quarters=8, through="2026Q2")
+print(panel.document["bank"]["name"], panel.document["coverage"])
+print(panel.dataset_id)
+```
+
+Omitting `through` asks FDIC for that bank's latest available reporting date and
+records that default in the manifest. Different banks can have different latest
+dates. Select the same explicit quarter when comparing banks. Requests are
+bounded and transient failures have capped retries.
+
+## Use your existing analysis tools
+
+```python
+frame = panel.to_pandas()       # optional pandas; evidence retained in frame.attrs["banksonsale"]
+records = panel.rows            # independent copy of standard Python records
+panel.export("live-oak.json")  # self-contained snapshot with archived evidence
+panel.export("live-oak.parquet")  # optional pyarrow; evidence in schema metadata
+```
+
+CSV works in spreadsheets and DuckDB. `.to_polars()` is available with Polars
+installed. See [the data contract](https://github.com/anshuman-pandey/banksonsale-data/blob/v0.1.0/docs/DATA_CONTRACT.md) for nulls, units and
+snapshot integrity. JSON from `panel.export()` is a BankPanel envelope; the
+website's quarter-wide JSON download is a separate documented format.
+
+## Run a real example
+
+- [Bank financials and exports](https://github.com/anshuman-pandey/banksonsale-data/blob/v0.1.0/notebooks/01_bank_financials.ipynb) — [Open in Colab](https://colab.research.google.com/github/anshuman-pandey/banksonsale-data/blob/v0.1.0/notebooks/01_bank_financials.ipynb).
+- [Same-quarter comparison and SQL](https://github.com/anshuman-pandey/banksonsale-data/blob/v0.1.0/notebooks/02_compare_banks.ipynb) — [Open in Colab](https://colab.research.google.com/github/anshuman-pandey/banksonsale-data/blob/v0.1.0/notebooks/02_compare_banks.ipynb).
+- [Ready-to-open comparison CSV](https://github.com/anshuman-pandey/banksonsale-data/blob/v0.1.0/examples/comparison-2026-q2.csv), with the [source manifest](https://github.com/anshuman-pandey/banksonsale-data/blob/v0.1.0/examples/comparison-2026-q2.manifest.json).
+
+The notebooks run locally without a cloud account. Their install cell retrieves
+the tagged public release. Colab has its own sign-in requirements and limits.
+
+## What you get
+
+Fourteen reported fields cover assets, total and domestic deposits, equity,
+intangibles, net loans, estimated uninsured deposits, core deposits, CET1,
+leverage, annualized ROA/ROE, domestic offices and employees.
+
+- Monetary columns end in `_usd` and are **U.S. dollars**, normalized once from FDIC's USD-thousands values.
+- Ratios end in `_pct`: **12.5 means 12.5%**. ROA and ROE are already annualized by the source.
+- CET1 uses **RBCT1CER**. Leverage uses **RBC1AAJ**. They remain distinct.
+- Missing observations stay `None`. Reported zero and negative observations remain unchanged.
+- Missing quarters remain explicit rows with `report_available=False`; older reports are never substituted.
+- FDIC CERT identifies an institution, not a listed holding company, ticker or individual branch.
+
+This is a small, versioned release of selected FDIC fields. It does not cover all
+bank data, customers' accounts, live transactions, full Call Reports, transaction
+prices or verified sale listings. Names and institution metadata reflect the
+upstream source at retrieval, not a reconstructed historical identity timeline.
+
+## Command line
+
+```bash
+banksonsale-data sample --cert 58665 --output live-oak.csv
+banksonsale-data search "Live Oak" --state NC
+banksonsale-data financials --cert 58665 --quarters 8 --through 2026Q2 --output live-oak.json
+banksonsale-data definitions
+```
+
+## Continue the research in BanksOnSale
+
+[Open Live Oak's public financial summary](https://www.banksonsale.com/banks/58665)
+or [explore the BanksOnSale research workspace](https://www.banksonsale.com/discover).
+The hosted product adds saved screens, watchlists, research dossiers, monitoring,
+models and exports; some features require a subscription. Its data edition is
+displayed separately and can differ from this package's selected period.
+
+These are optional links. The package keeps working without an account, sends
+no usage telemetry, and does not create a subscription. See [continuing your
+research](https://github.com/anshuman-pandey/banksonsale-data/blob/v0.1.0/docs/CONTINUE_IN_BANKSONSALE.md).
+
+## OpenBB
+
+The optional [independently maintained OpenBB V5 extension](https://github.com/anshuman-pandey/banksonsale-data/tree/v0.1.0/integrations/openbb)
+adds `obb.banksonsale`. It reuses this package and its definitions. It is not
+bundled with or endorsed by OpenBB. The core package does not require OpenBB.
+
+## Source, license and citation
+
+Original source: [FDIC BankFind Suite](https://api.fdic.gov/banks/docs/).
+BanksOnSale is an independent compiler; no FDIC affiliation or endorsement is implied.
+
+Original code: [MIT](https://github.com/anshuman-pandey/banksonsale-data/blob/v0.1.0/LICENSE). Sample compilation and original metadata:
+[CC BY 4.0, with separate source attribution](https://github.com/anshuman-pandey/banksonsale-data/blob/v0.1.0/DATA_LICENSE.md).
+Use [CITATION.cff](https://github.com/anshuman-pandey/banksonsale-data/blob/v0.1.0/CITATION.cff) and record the bank CERT, report dates and dataset ID.
+SHA-256 checks detect changes to a snapshot; they are not a government signature
+or proof that a filing is economically correct.
+
+## Contributing and verification
+
+See [CONTRIBUTING.md](https://github.com/anshuman-pandey/banksonsale-data/blob/v0.1.0/CONTRIBUTING.md). Offline checks:
+
+```bash
+python -m pip install -e .
+python -m unittest discover -s tests -v
+```
+
+Network refreshes are explicit maintainer actions, not part of the default tests.
+[Report a package issue](https://github.com/anshuman-pandey/banksonsale-data/issues)
+or [a source-data correction](https://www.banksonsale.com/corrections).

banksonsale_data-0.1.0/docs/CONTINUE_IN_BANKSONSALE.md

@@ -0,0 +1,25 @@
+# Continue research in BanksOnSale
+
+The open package works independently of a BanksOnSale subscription. You can keep
+its samples, notebook results and source-linked exports without registering.
+
+For a selected institution, open its public summary using its FDIC CERT:
+
+```python
+url = f"https://www.banksonsale.com/banks/{panel.document['bank']['cert']}"
+print(url)
+```
+
+The public summary links to the bank's guarded research dossier. The hosted
+workspace supports saved screens, watchlists, research dossiers, monitoring,
+models and supported exports. Some features require Professional access. Current
+plans and trial terms are at https://www.banksonsale.com/pricing.
+
+The hosted workspace displays its own financial reporting date and release.
+That edition can differ from this package's selected quarter. A bank identity
+link is not a claim that the chosen period or every notebook filter will survive
+an upgrade/checkout flow. Check the displayed coverage before continuing a
+comparison; do not silently substitute a different bank or period.
+
+Links in the package are optional and have no hidden telemetry. This library
+never sends names, emails, saved research or usage events to BanksOnSale.

banksonsale_data-0.1.0/docs/DATA_CONTRACT.md

@@ -0,0 +1,55 @@
+# BanksOnSale Data contract 1.0
+
+`BankPanel` contains one explicitly selected FDIC institution and a contiguous
+calendar-quarter window of 1–40 quarters. A report can be unavailable within the
+window. Its row remains present with `report_available=false` and null metrics;
+no prior or subsequent quarter is substituted.
+
+## Identity and time
+
+- `bank.cert`: FDIC CERT. This identifies an institution, not its listed parent, ticker or branch.
+- `bank.rssd_id`: the source's RSSD ID, when reported.
+- `bank.name`, city, state and activity metadata: institution attributes at retrieval. These are not a historical identity reconstruction. Consult `bank_class` when available; the API can cover different FDIC institution classes.
+- `requested_dates` and `rows[].report_date`: exact financial reporting quarter ends.
+- `sources[].retrieved_at`: UTC time this package fetched the response.
+- `sources[].source_index_created_at`: FDIC index creation timestamp, not the report date.
+- `report_date_defaulted`: true only when the caller omitted `through` and the client selected that bank's latest available report date.
+
+## Numbers
+
+Money is normalized from FDIC USD thousands to **USD** exactly once. Ratio
+observations retain **percent** units: 12.5 is 12.5%, not 0.125%. ROA/ROE are
+annualized by the source and never annualized again. The definition for each
+column records its raw field, multiplier, reporting basis and documentation URL.
+CET1 is RBCT1CER; leverage is RBC1AAJ. Equity/assets is a separate derived
+accounting ratio and must not be relabeled as either regulatory ratio.
+
+Blank/NA/absent observations become null. Reported zeros and negative values are
+retained. Missing/inapplicable values are not inferred from nearby observations.
+Malformed, boolean, infinite or nonnumeric financial values fail validation.
+A reported zero's economic applicability is not inferred.
+
+## Evidence and portability
+
+Each panel carries the selected field definitions, coverage counts, exact FDIC
+request URLs, retrieved raw JSON, source-response SHA-256 hashes, a content hash
+and a branded dataset ID. Integrity checks detect byte/content changes; they are
+not a digital signature from FDIC or independent verification of a filing.
+
+`panel.export('file.json')` saves the self-contained envelope. Reload it with
+`BankPanel.load('file.json')`. The website's all-institution quarter JSON is a
+separate format; do not pass its raw array to `BankPanel.load()`.
+
+CSV exports include a matching `.manifest.json` file carrying evidence, units,
+coverage and a SHA-256 of the CSV bytes. Keep those files together. Parquet embeds
+the manifest in schema metadata under `banksonsale`. pandas uses
+`frame.attrs['banksonsale']`. Polars users should retain `panel.manifest()` or
+export a self-contained artifact alongside the DataFrame.
+
+## Scope
+
+The current release exposes 14 selected FDIC financial fields. It does not
+provide full Call Report/XBRL coverage, customer-level data, real-time balances,
+complete ownership history, transaction valuations or verified sale intent.
+Current upstream revisions are not information-as-known-at-the-time history.
+This is research software, not an investment, regulatory or safety rating.

banksonsale_data-0.1.0/docs/METRICS.md

@@ -0,0 +1,24 @@
+# BanksOnSale Data metric definitions
+
+The package normalizes source monetary observations to USD. The website's raw FDIC downloads retain USD thousands; check the format before joining.
+
+| Package column | FDIC field | Package unit | Multiplier | Basis |
+| --- | --- | --- | --- | --- |
+| `assets_usd` | `ASSET` | USD | 1000 | period_end |
+| `deposits_usd` | `DEP` | USD | 1000 | period_end |
+| `domestic_deposits_usd` | `DEPDOM` | USD | 1000 | period_end |
+| `equity_usd` | `EQ` | USD | 1000 | period_end |
+| `intangibles_usd` | `INTAN` | USD | 1000 | period_end |
+| `net_loans_usd` | `LNLSNET` | USD | 1000 | period_end |
+| `uninsured_deposits_usd` | `DEPUNINS` | USD | 1000 | period_end |
+| `core_deposits_usd` | `COREDEP` | USD | 1000 | period_end |
+| `cet1_ratio_pct` | `RBCT1CER` | percent | 1 | period_end |
+| `leverage_ratio_pct` | `RBC1AAJ` | percent | 1 | period_end |
+| `roa_pct` | `ROA` | percent | 1 | annualized_by_source |
+| `roe_pct` | `ROE` | percent | 1 | annualized_by_source |
+| `domestic_offices` | `OFFDOM` | count | 1 | period_end |
+| `employees` | `NUMEMP` | count | 1 | period_end |
+
+Original definitions: https://api.fdic.gov/banks/docs/risview_properties.yaml
+
+Null observations are never converted to zero. Ratio values are already percentages. See [the data contract](DATA_CONTRACT.md) for time and identity rules.

banksonsale_data-0.1.0/docs/PUBLISHING.md

@@ -0,0 +1,34 @@
+# Publishing BanksOnSale Data
+
+The GitHub release is independently installable. A PyPI release is a separate
+publication and must not be advertised as available before its project page and
+clean index installation actually work.
+
+## PyPI account setup
+
+The repository has a manual `publish-pypi.yml` workflow using PyPI's trusted
+publishing protocol; no API token belongs in source, issues, chat or workflows.
+Before running it, the account owner must register the publisher on PyPI for both
+`banksonsale-data` and `openbb-banksonsale`:
+
+- Repository owner: `anshuman-pandey`
+- Repository name: `banksonsale-data`
+- Workflow filename: `publish-pypi.yml`
+- Environment name: `pypi`
+
+For an unpublished package, use PyPI's pending-publisher flow. Account creation,
+2FA and new publishing trust require the account owner's action/authorization.
+Reference: https://docs.pypi.org/trusted-publishers/creating-a-project-through-oidc/
+
+After registration, run the manual GitHub workflow against the tested tag. It
+rebuilds and checks the core plus OpenBB distributions and publishes with short-
+lived OIDC credentials. Verify both PyPI project pages and clean installations
+before changing README/website instructions to registry shorthand.
+
+## Release integrity
+
+Build and run offline tests, notebook execution, installed-wheel checks and the
+optional OpenBB integration before tagging. Pin source data and retain its
+manifest; no silent replacement of sample bytes under an existing release tag.
+Upload checksummed artifacts to the GitHub release. Never bundle private website
+source, user records, prospect lists, API keys or account exports.