sdm-learn 0.1.0.dev0__tar.gz

sdm_learn-0.1.0.dev0/LICENSE

@@ -0,0 +1,201 @@
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "[]"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright [yyyy] [name of copyright owner]
+
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.

sdm_learn-0.1.0.dev0/MANIFEST.in

@@ -0,0 +1,10 @@
+include LICENSE
+include README.md
+recursive-include examples *.py
+recursive-include tests *.py
+
+# The GIFs stay out of the distribution: 2.5 MB of animation helps a reader on
+# GitHub and does nothing for an install. The README points at them by URL so
+# the PyPI page still shows them.
+prune assets
+prune runs

sdm_learn-0.1.0.dev0/PKG-INFO

@@ -0,0 +1,35 @@
+Metadata-Version: 2.4
+Name: sdm-learn
+Version: 0.1.0.dev0
+Summary: Self-documenting models: the document is the model.
+Author: Yulu Gan
+License-Expression: Apache-2.0
+Project-URL: Homepage, https://github.com/sunrainyg/sdm
+Project-URL: Repository, https://github.com/sunrainyg/sdm
+Keywords: machine-learning,llm,interpretability,self-documenting,supervised-learning
+Classifier: Development Status :: 4 - Beta
+Classifier: Intended Audience :: Science/Research
+Classifier: Programming Language :: Python :: 3
+Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
+Classifier: Operating System :: OS Independent
+Requires-Python: >=3.9
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: httpx>=0.24
+Provides-Extra: image
+Requires-Dist: Pillow>=9; extra == "image"
+Requires-Dist: cairosvg>=2.5; extra == "image"
+Provides-Extra: audio
+Requires-Dist: numpy>=1.21; extra == "audio"
+Requires-Dist: Pillow>=9; extra == "audio"
+Provides-Extra: video
+Requires-Dist: numpy>=1.21; extra == "video"
+Requires-Dist: Pillow>=9; extra == "video"
+Requires-Dist: cairosvg>=2.5; extra == "video"
+Provides-Extra: all
+Requires-Dist: numpy>=1.21; extra == "all"
+Requires-Dist: Pillow>=9; extra == "all"
+Requires-Dist: cairosvg>=2.5; extra == "all"
+Dynamic: license-file
+
+# sdm

sdm_learn-0.1.0.dev0/README.md

@@ -0,0 +1 @@
+# sdm

sdm_learn-0.1.0.dev0/examples/demo.py

@@ -0,0 +1,319 @@
+"""A runnable demonstration of supervised learning in a skill document.
+
+    python demo.py --fake                  # no API key needed
+    python demo.py --fake --format html    # the same loop, HTML skill
+    python demo.py                         # against the real model
+
+It scores the no-document baseline first, then learns a skill document from the
+training rows under the gate and scores again, so the printed accuracies
+bracket whatever the learned document was worth.
+
+``--fake`` swaps the model for a keyword stub that genuinely reads its own
+skill document and genuinely rewrites it from the revealed labels. It is far
+weaker than an LLM, but it exercises the same loop, so the demo runs offline
+and still shows accuracy moving as the document accumulates signatures.
+"""
+
+from __future__ import annotations
+
+import argparse
+import re
+from collections import Counter
+
+import sdm
+from sdm import gateway
+
+
+# --------------------------------------------------------------------------
+# A small labeled dataset. No row contains its own label as a word.
+# --------------------------------------------------------------------------
+
+TRAIN = [
+    ("Congratulations, you are a winner! Claim your free prize now.", "junk"),
+    ("Standup moved to 10am, check the calendar invite.", "real"),
+    ("URGENT: wire the deposit today or the offer expires.", "junk"),
+    ("Here are the slides from yesterday's review.", "real"),
+    ("Claim your free cash bonus, click the link to continue.", "junk"),
+    ("Can you review my pull request when you get a chance?", "real"),
+    ("Guaranteed income from home, no experience needed, act now.", "junk"),
+    ("The build is failing on main and the tests are red.", "real"),
+    ("You have won a free cruise! Click here to claim the prize.", "junk"),
+    ("Moving our one-on-one to Thursday, does that work?", "real"),
+    ("Final notice: wire transfer required to release your funds.", "junk"),
+    ("I pushed the fix, the tests are green now.", "real"),
+    ("Exclusive offer, guaranteed returns, double your cash today.", "junk"),
+    ("Left a few comments on the design doc.", "real"),
+    ("Free gift card waiting, claim it before the offer ends.", "junk"),
+    ("Can we push the deploy until after the demo?", "real"),
+    ("Lucky winner! Click to claim your cash prize now.", "junk"),
+    ("Adding the review notes to the calendar invite.", "real"),
+    ("Verify your account immediately or lose access forever.", "junk"),
+    ("The deploy is blocked until the tests pass.", "real"),
+]
+
+TEST = [
+    ("Claim your free gift card, the offer expires tonight.", "junk"),
+    ("Do you have the slides from the review meeting?", "real"),
+    ("Guaranteed cash prize, click the link to claim.", "junk"),
+    ("Rescheduling the standup to Friday, see the calendar.", "real"),
+    ("URGENT: wire the funds today, this offer is final.", "junk"),
+    ("I left comments on the pull request, the tests are green.", "real"),
+    ("You are our lucky winner, claim your free bonus.", "junk"),
+    ("The deploy is waiting on the failing build.", "real"),
+]
+
+CLASSES = ("junk", "real")
+DESCRIPTION = "Each input is the body of one email."
+
+
+def build() -> sdm.Dataset:
+    return sdm.dataset(
+        [{"text": text, "label": label} for text, label in TRAIN],
+        [{"text": text, "label": label} for text, label in TEST],
+        classes=CLASSES, name="email", title="Email skill",
+        description=DESCRIPTION,
+    )
+
+
+# --------------------------------------------------------------------------
+# The offline stub
+# --------------------------------------------------------------------------
+
+STOPWORDS = {
+    "the", "a", "an", "and", "or", "to", "of", "in", "on", "for", "your",
+    "you", "i", "is", "are", "was", "it", "me", "my", "we", "at", "that",
+    "this", "with", "from", "have", "has", "will", "be", "not", "no", "if",
+    "can", "do", "does", "now", "today", "input", "appears", "when",
+    # the stub's own document phrasing, so it never counts as evidence
+    "body", "mentions", "wording", "such", "any", "one", "signature",
+    "signatures", "email", "skill", "learned", "class", "doctype", "html",
+    "head", "meta", "charset", "utf", "title", "li", "ul", "hr",
+}
+RULE_LINE_RE = re.compile(r"(R\d{3}) \[support=\d+ used=\d+ hit=\d+\] (.+)")
+LIST_ITEM_RE = re.compile(r"<li>\s*(\w+):\s*([^<]+)</li>")
+SKILL_FILE_RE = re.compile(r"<skill_file[^>]*>(.*?)</skill_file>", re.DOTALL)
+SOURCE_RE = re.compile(r"<current_source>(.*?)</current_source>", re.DOTALL)
+EXAMPLE_RE = re.compile(
+    r'<example id="P\d+" true="([^"]+)"[^>]*>\s*(.*?)\s*<evidence>',
+    re.DOTALL)
+INPUT_RE = re.compile(r'<input id="P(\d+)">\s*(.*?)\s*</input>', re.DOTALL)
+CURRENT_RE = re.compile(
+    r'<current_input id="P(\d+)">\s*(.*?)\s*</current_input>', re.DOTALL)
+DOCUMENT_RE = re.compile(r'<document id="([AB])">(.*?)</document>', re.DOTALL)
+BLOCK_RE = re.compile(r"[Rr]eturn (?:ONLY|only) an? <(\w+)> block")
+
+
+def words(text: str) -> set[str]:
+    return {word for word in re.findall(r"[a-z]+", text.lower())
+            if len(word) > 2 and word not in STOPWORDS}
+
+
+def read_document(text: str) -> dict[str, tuple[str, set[str]]]:
+    """Per-class (rule id, signature words), from Markdown or from HTML.
+
+    The stub keeps no memory of its own. Like the model, it reads back the
+    document it wrote last time, whatever format that document is in.
+    """
+    found: dict[str, tuple[str, set[str]]] = {}
+    label = "general"
+    for line in text.splitlines():
+        if line.startswith("### "):
+            label = line[4:].strip()
+        match = RULE_LINE_RE.search(line)
+        if match and label in CLASSES:
+            found[label] = (match.group(1), words(match.group(2)))
+    for name, listed in LIST_ITEM_RE.findall(text):
+        if name in CLASSES:
+            found[name] = ("none", words(listed))
+    return found
+
+
+def flatten(messages: list) -> str:
+    parts = []
+    for message in messages:
+        body = message["content"]
+        parts.append(body if isinstance(body, str) else "\n".join(
+            block["text"] for block in body if block.get("type") == "text"))
+    return "\n".join(parts)
+
+
+class KeywordModel:
+    """A stub that follows the document it is given and then rewrites it."""
+
+    def __init__(self):
+        self.revisions = 0
+
+    def _classify(self, text: str, learned) -> tuple[str, str]:
+        best, best_score, best_id = CLASSES[0], 0, "none"
+        for label, (rule_id, signature) in learned.items():
+            score = len(signature & words(text))
+            if score > best_score:
+                best, best_score, best_id = label, score, rule_id
+        return best, best_id
+
+    def _predictions(self, prompt: str) -> str:
+        block = SKILL_FILE_RE.search(prompt)
+        learned = read_document(block.group(1) if block else "")
+        found = INPUT_RE.findall(prompt) or CURRENT_RE.findall(prompt)
+        cites = "R001,R002 or none" in prompt
+        lines = []
+        for number, text in found:
+            label, rule_id = self._classify(text, learned)
+            middle = f" {rule_id} |" if cites else ""
+            lines.append(f"P{int(number):02d} | {label} |{middle} "
+                         f"keyword overlap")
+        return "<predictions>\n" + "\n".join(lines) + "\n</predictions>"
+
+    def _gate(self, prompt: str) -> str:
+        tail = prompt.split("</document>")[-1]
+        found = CURRENT_RE.findall(tail) or INPUT_RE.findall(tail)
+        blocks = []
+        for letter, body in DOCUMENT_RE.findall(prompt):
+            learned = read_document(body)
+            lines = [f"P{int(number):02d} | "
+                     f"{self._classify(text, learned)[0]} | keyword overlap"
+                     for number, text in found]
+            blocks.append(f'<predictions document="{letter}">\n'
+                          + "\n".join(lines) + "\n</predictions>")
+        return "\n".join(blocks)
+
+    def _signatures(self, prompt: str) -> dict[str, list[str]]:
+        """Grow each class signature with words this batch separated cleanly."""
+        source = SOURCE_RE.search(prompt)
+        learned = read_document(source.group(1) if source else "")
+        known = {word for _, signature in learned.values()
+                 for word in signature}
+        counts: dict[str, Counter] = {label: Counter() for label in CLASSES}
+        for label, body in EXAMPLE_RE.findall(prompt):
+            if label in counts:
+                counts[label].update(words(body) - known)
+
+        grown: dict[str, list[str]] = {}
+        for label in CLASSES:
+            others = Counter()
+            for other, counter in counts.items():
+                if other != label:
+                    others.update(counter)
+            fresh = sorted(word for word, count in counts[label].items()
+                           if not others[word])[:6]
+            existing = learned.get(label, ("none", set()))[1]
+            if fresh or existing:
+                grown[label] = sorted(existing | set(fresh))[:30]
+        return grown
+
+    def _revision(self, prompt: str) -> str:
+        self.revisions += 1
+        block = BLOCK_RE.search(prompt)
+        name = block.group(1) if block else "edits"
+        grown = self._signatures(prompt)
+        if name == "edits":
+            return self._edits(prompt, grown)
+        return f"<{name}>\n{self._page(grown)}\n</{name}>"
+
+    @staticmethod
+    def _edits(prompt: str, grown: dict[str, list[str]]) -> str:
+        source = SOURCE_RE.search(prompt)
+        learned = read_document(source.group(1) if source else "")
+        edits = []
+        for label, signature in grown.items():
+            listed = ", ".join(signature)
+            if label in learned and learned[label][0] != "none":
+                edits.append(f"REVISE | {learned[label][0]} | wording such "
+                             f"as {listed}")
+            else:
+                edits.append(f"ADD | {label} | wording such as {listed}")
+        return "<edits>\n" + "\n".join(edits or ["NOOP"]) + "\n</edits>"
+
+    @staticmethod
+    def _page(grown: dict[str, list[str]]) -> str:
+        """The same learned signatures, written as a standalone HTML page."""
+        items = "\n".join(f"<li>{label}: {', '.join(signature)}</li>"
+                          for label, signature in grown.items())
+        return f"""<!DOCTYPE html>
+<html>
+<head><meta charset="utf-8"><title>Email skill</title></head>
+<body>
+<h1>Email skill</h1>
+<h2>Reading procedure</h2>
+<p>Score the input against each signature below and pick the best overlap.</p>
+<h2>Per-class signatures</h2>
+<ul>
+{items}
+</ul>
+</body>
+</html>"""
+
+    def __call__(self, session, messages, max_tokens, effort=None) -> str:
+        prompt = flatten(messages)
+        if "deterministic evaluation component" in prompt:
+            return self._gate(prompt)
+        if "candidate revision" in prompt:
+            return self._revision(prompt)
+        return self._predictions(prompt)
+
+
+class Session:
+    def close(self):
+        pass
+
+
+def use_stub() -> None:
+    # Every module reaches the gateway through model, so replacing the
+    # transport there is enough to guarantee this run stays offline.
+    gateway.complete = KeywordModel()
+    gateway.client = lambda *args, **kwargs: Session()
+
+
+# --------------------------------------------------------------------------
+
+def main(argv=None) -> int:
+    parser = argparse.ArgumentParser(description=__doc__)
+    parser.add_argument("--fake", action="store_true",
+                        help="use the offline keyword stub, no API key needed")
+    parser.add_argument("--format", default=None,
+                        help="a section of format.md; the stub can drive "
+                             "'markdown' and 'html'")
+    parser.add_argument("--batch-size", type=int, default=4)
+    parser.add_argument("--cycles", type=int, default=2)
+    parser.add_argument("--run-dir", default=None,
+                        help="checkpoint the run here")
+    arguments = parser.parse_args(argv)
+
+    if arguments.fake:
+        use_stub()
+        print("FAKE MODEL: a keyword stub stands in for the LLM. The loop is\n"
+              "real, the accuracies are only as good as the stub.\n")
+
+    data = build()
+    chosen = arguments.format or sdm.load_formats()[0]
+    print(f"{data}\nskill format: {chosen}\n")
+
+    try:
+        control = sdm.train(data, algo="baseline")
+        empty = sdm.evaluate(control, data, batch_size=arguments.batch_size)
+        print(f"\nno document:      {empty}")
+
+        skill = sdm.train(data, algo="gate", format=arguments.format,
+                           batch_size=arguments.batch_size,
+                           holdout_per_class=2, cycles=arguments.cycles,
+                           run_dir=arguments.run_dir)
+        learned = sdm.evaluate(skill, data, batch_size=arguments.batch_size)
+        print(f"\nlearned document: {learned}")
+    except sdm.SetupNeeded as error:
+        print(f"\n{error}\n\nOr run this demo offline with --fake.")
+        return 1
+
+    accepted = sum(entry["decision"]["accepted"]
+                   for entry in skill.state["candidates"])
+    print(f"\ndelta: {learned.accuracy - empty.accuracy:+.1%} on "
+          f"{learned.total} test rows, "
+          f"{accepted}/{len(skill.state['candidates'])} candidates accepted, "
+          f"{len(skill.state['request_audit'])} requests spent")
+    print(f"\n{'-' * 70}\n{skill.text}\n{'-' * 70}")
+    if arguments.run_dir:
+        print(f"\nrun saved to {arguments.run_dir}/")
+    return 0
+
+
+if __name__ == "__main__":
+    raise SystemExit(main())

sdm_learn-0.1.0.dev0/pyproject.toml

@@ -0,0 +1,57 @@
+[build-system]
+requires = ["setuptools>=64"]
+build-backend = "setuptools.build_meta"
+
+[project]
+# The distribution name is sdm-learn because PyPI's `sdm` is taken by an
+# unrelated 2018 package. The import name is `sdm`, which is what appears in
+# the paper and in every example; scikit-learn/sklearn is the same split.
+name = "sdm-learn"
+# One source of truth: sdm.__version__. A PyPI version number can never be
+# reused, so a stale duplicate here is the kind of mistake you cannot undo.
+dynamic = ["version"]
+description = "Self-documenting models: the document is the model."
+readme = "README.md"
+requires-python = ">=3.9"
+license = "Apache-2.0"
+license-files = ["LICENSE"]
+authors = [{ name = "Yulu Gan" }]
+keywords = ["machine-learning", "llm", "interpretability",
+            "self-documenting", "supervised-learning"]
+classifiers = [
+    "Development Status :: 4 - Beta",
+    "Intended Audience :: Science/Research",
+    "Programming Language :: Python :: 3",
+    "Topic :: Scientific/Engineering :: Artificial Intelligence",
+    "Operating System :: OS Independent",
+]
+dependencies = ["httpx>=0.24"]
+
+[project.optional-dependencies]
+# One extra per skill format that renders something. A text-only install needs
+# none of these, which is why they are not in `dependencies`.
+image = ["Pillow>=9", "cairosvg>=2.5"]
+audio = ["numpy>=1.21", "Pillow>=9"]
+video = ["numpy>=1.21", "Pillow>=9", "cairosvg>=2.5"]
+all = ["numpy>=1.21", "Pillow>=9", "cairosvg>=2.5"]
+
+[project.scripts]
+sdm = "sdm.__main__:main"
+
+[project.urls]
+Homepage = "https://github.com/sunrainyg/sdm"
+Repository = "https://github.com/sunrainyg/sdm"
+
+[tool.setuptools]
+package-dir = { "" = "src" }
+
+[tool.setuptools.dynamic]
+version = { attr = "sdm.__version__" }
+
+[tool.setuptools.packages.find]
+where = ["src"]
+
+[tool.setuptools.package-data]
+# format.md ships with the package as the read-only default. `sdm init` copies
+# it out so a project can edit its own.
+sdm = ["format.md"]

sdm_learn-0.1.0.dev0/setup.cfg

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