gruplac-json 0.1.0__py3-none-any.whl

gruplac_json/__init__.py

@@ -0,0 +1,17 @@
+"""gruplac-json: turn a public GrupLAC page (Minciencias, Colombia) into one JSON document.
+
+>>> from gruplac_json import extract
+>>> data = extract("COL0008423")            # downloads the page; Scienti ids if an apis.colav.co key is set
+>>> data["name"], len(data["members"]), len(data["works"])
+"""
+
+from importlib.metadata import PackageNotFoundError, version
+
+try:
+    __version__ = version("gruplac-json")
+except PackageNotFoundError:  # running from a source tree that was never installed
+    __version__ = "0.0.0"
+
+from .parser import build, extract, fetch, resolve
+
+__all__ = ["__version__", "build", "extract", "fetch", "resolve"]

gruplac_json/__main__.py

@@ -0,0 +1,3 @@
+from .cli import main
+
+raise SystemExit(main())

gruplac_json/cli.py

@@ -0,0 +1,68 @@
+"""Command line entry point: `gruplac-json`."""
+
+import argparse
+import gzip
+import json
+import sys
+
+from . import __version__
+from .parser import decode, extract, fetch, resolve
+
+
+def main(argv=None):
+    ap = argparse.ArgumentParser(
+        prog="gruplac-json",
+        description="Build a JSON from a research group's public GrupLAC page (Minciencias, Colombia)",
+    )
+    ap.add_argument("group", help="COL code (COL0008423) or GrupLAC number (00000000001676; leading zeros optional)")
+    ap.add_argument("-o", "--output", help="output file (default gruplac_<COL or number>.json; '-' for stdout)")
+    ap.add_argument("--html", help="parse a saved page (.html or .html.gz) instead of downloading it")
+    ap.add_argument("--save-html", help="also save the downloaded page here")
+    ap.add_argument("--scienti", choices=["auto", "off"], default="auto", help="add Scienti ids (needs an apis.colav.co key)")
+    ap.add_argument("--institution", help="Scienti dump (udea, univalle, uec, unaula); default: try them in turn")
+    ap.add_argument("--model-year", help="Scienti model year (default: the dump's latest)")
+    ap.add_argument("--apikey", help="apis.colav.co key (else $COLAV_APIKEY or ~/.config/colav/apikey)")
+    ap.add_argument("--ror", choices=["auto", "off"], default="auto", help="look institutions up in the ROR registry")
+    ap.add_argument("--version", action="version", version=f"%(prog)s {__version__}")
+    a = ap.parse_args(argv)
+    log = lambda *x: print(*x, file=sys.stderr)
+    try:
+        nro, col = resolve(a.group)
+    except ValueError as e:
+        ap.error(str(e))
+    if a.html:
+        with open(a.html, "rb") as fh:
+            raw = fh.read()
+        html = decode(gzip.decompress(raw) if raw[:2] == b"\x1f\x8b" else raw)
+    else:
+        html = fetch(nro)
+        if a.save_html:
+            with open(a.save_html, "w", encoding="latin-1", errors="xmlcharrefreplace") as fh:  # as the page declares
+                fh.write(html)
+    data = extract(
+        col or nro,
+        html=html,
+        scienti=a.scienti,
+        institution=a.institution,
+        model_year=a.model_year,
+        apikey=a.apikey,
+        ror=a.ror == "auto",
+        log=log,
+    )
+    out = a.output or f"gruplac_{col or nro}.json"
+    text = json.dumps(data, ensure_ascii=False, indent=1)
+    if out == "-":
+        print(text)
+    else:
+        with open(out, "w", encoding="utf-8") as fh:
+            fh.write(text + "\n")
+        c = data["counts"]
+        log(
+            f"{out}: {data['name']} - {c['members']} members, {c['works']} works, {c['projects']} projects, "
+            f"{c['events']} events, {sum(len(v) for v in data['other_products'].values())} other products"
+        )
+    return 0
+
+
+if __name__ == "__main__":
+    raise SystemExit(main())