gismap 0.1.0__py3-none-any.whl → 0.2.0__py3-none-any.whl

gismap/utils/requests.py

@@ -1,12 +1,19 @@
 from time import sleep
 import requests
+from importlib.metadata import metadata
 from gismap.utils.logger import logger
 
 
+infos = metadata("gismap")
 session = requests.Session()
+session.headers.update(
+    {
+        "User-Agent": f"{infos['name']}/{infos['Version']} ({'; '.join(infos.get_all('Project-URL'))}; Contact, {infos['author-email']}"
+    }
+)
 
 
-def get(url, params=None):
+def get(url, params=None, n_trials=10):
     """
     Parameters
     ----------
@@ -20,14 +27,20 @@ def get(url, params=None):
     :class:`str`
         Result.
     """
-    while True:
-        r = session.get(url, params=params)
-        if r.status_code == 429:
-            try:
-                t = int(r.headers["Retry-After"])
-            except KeyError:
-                t = 60
-            logger.warning(f"Too many requests. Auto-retry in {t} seconds.")
+    for attempt in range(n_trials):
+        try:
+            r = session.get(url, params=params)
+            if r.status_code == 429:
+                try:
+                    t = int(r.headers["Retry-After"])
+                except KeyError:
+                    t = 60
+                logger.warning(f"Too many requests. Auto-retry in {t} seconds.")
+                sleep(t)
+            else:
+                return r.text
+        except requests.exceptions.ConnectionError:
+            t = 6
+            logger.warning(f"Connection error. Auto-retry in {t} seconds.")
             sleep(t)
-        else:
-            return r.text
+    raise requests.exceptions.ConnectionError(f"Unable to retrieve {url}")

gismap/utils/text.py

@@ -1,4 +1,5 @@
 from bof.fuzz import Process
+import unicodedata
 
 
 class Corrector:
@@ -90,4 +91,68 @@ def clean_aliases(name, alias_list):
     :class:`list`
         Aliases deduped, sorted, and with main name removed.
     """
-    return sorted(set(n for n in alias_list if n != name))
+    return sorted(set(n for nn in alias_list for n in [nn, asciify(nn)] if n != name))
+
+
+def asciify(text):
+    """
+    Parameters
+    ----------
+    text::class:`str`
+        Some text (typically names) with annoying accents.
+
+    Returns
+    -------
+    :class:`str`
+        Same text simplified into ascii.
+
+    Examples
+    --------
+    >>> asciify('Ana Bušić')
+    'Ana Busic'
+    >>> asciify("Thomas Deiß")
+    'Thomas Deiss'
+    """
+    text = text.replace("ß", "ss")
+    decomposed = unicodedata.normalize("NFD", text)
+    no_accents = "".join(c for c in decomposed if unicodedata.category(c) != "Mn")
+    ascii_text = no_accents.encode("ascii", "ignore").decode()
+    return ascii_text
+
+
+def normalized_name(txt):
+    """
+    Try to normalize names for facilitating comparisons. Name is lowered, split, asciified, sorted, and filtered.
+
+    Parameters
+    ----------
+    txt: :class:`str`
+
+    Returns
+    -------
+    :class:`str`
+
+    Examples
+    --------
+
+    >>> normalized_name("Thomas Deiß")
+    'deiss thomas'
+    >>> normalized_name("Dario Rossi 001")
+    'dario rossi'
+    >>> normalized_name("James W. Roberts")
+    'james roberts'
+    """
+    return " ".join(
+        sorted(
+            asciify(a)
+            for a in txt.lower().replace("-", " ").split()
+            if not (a.isdigit() or (len(a) < 3 and "." in a))
+        )
+    )
+
+
+def auto_int(txt):
+    try:
+        return int(txt)
+    except ValueError:
+        return txt

{gismap-0.1.0.dist-info → gismap-0.2.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: gismap
-Version: 0.1.0
+Version: 0.2.0
 Summary: GISMAP leverages DBLP and HAL databases to provide cartography tools for you and your lab.
 Project-URL: Repository, https://github.com/balouf/gismap
 Project-URL: Documentation, https://balouf.github.io/gismap
@@ -11,6 +11,8 @@ License-File: AUTHORS.md
 Requires-Python: >=3.10
 Requires-Dist: bof>=0.3.5
 Requires-Dist: gismo>=0.5.2
+Requires-Dist: ipykernel>=6.30.1
+Requires-Dist: tqdm>=4.67.1
 Description-Content-Type: text/markdown
 
 # Generic Information Search: Mapping and Analysis of Publications

gismap-0.2.0.dist-info/RECORD

@@ -0,0 +1,29 @@
+gismap/__init__.py,sha256=wZhnsqnlouh2Ugsb4L36K9QG8IEGrvI7s9biEv6LjRw,735
+gismap/author.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+gismap/gismap.py,sha256=h0hwdogXGFqerm-5ZPeT-irPn91pCcQRjiHThXsRzEk,19
+gismap/gismo.py,sha256=kBzvlo-fqcvEgwQt-c_hHkfI1oTjBbKKb54vM8il_nI,6295
+gismap/search.py,sha256=nsUoDsFGeEtvCZ0dB7ooRPC_6qsazkiWx_oM7dHdNV4,4932
+gismap/lab/__init__.py,sha256=WeJkoNZmFnLoEgK_nwXqZrA4nQtpeTLPf6PPMFIGb4k,415
+gismap/lab/egomap.py,sha256=L5OKeZR7RCfX3EPm859-jZ1jnqlC3kpa0jybofhLPPc,1328
+gismap/lab/expansion.py,sha256=fYcUvkGi3W1GTx2gnFT1z3cI-jlZFFAicDM99wi2wzE,6186
+gismap/lab/filters.py,sha256=LL3jeCqLBpLeaz1lz5IvduV4G1xBPaty9SHiBxoYgak,486
+gismap/lab/graph.py,sha256=UTmuJsUXC9ntmAWoH46cBW3CTIGKus5z1G045yBLzAQ,6673
+gismap/lab/lab.py,sha256=Uc9Qg7Okf3L5K8gHTlnOv7lPbeA9OBfivCmZwRPwzp8,5331
+gismap/lab/lab_author.py,sha256=jQhq1spZP5P-Bpzr5E_OP73gWy3SG6UjiVCGotjm57s,2439
+gismap/lab/lip6.py,sha256=LLLjV8zNqwdkvtuV4nfmScCDlaeeGw79I0KQuOQ1_Yo,1443
+gismap/lab/toulouse.py,sha256=Ex3lBPdYhAzAnPQw9Driho7joJNTHLpE_nNMooQH_nc,1495
+gismap/lab/vis.py,sha256=j92WFmYZ0Szg10tq332PT5jQn3YbJSIOkYraVcmxiNk,8293
+gismap/sources/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+gismap/sources/dblp.py,sha256=_AgPgDbHYYaeJVam43eExv1VsglvAV46eP77RL7o9Fo,5278
+gismap/sources/hal.py,sha256=3VjiUH5amNfJYV5uAtX-JcGWMNKUkmjfXU0Osca9CC8,10021
+gismap/sources/models.py,sha256=yJPBXcJO6MgOrSXgpGPepHDDtniJP1OsK8OSz__VzYc,543
+gismap/sources/multi.py,sha256=HQMQ4kcDwg0rZ42av98DZ17o83dpR2fD4KtP1yFX14c,4352
+gismap/utils/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+gismap/utils/common.py,sha256=nx1f60yNwFpl1oz08h-R5o0xK9CbJv9tmYLDk61dwYA,2898
+gismap/utils/logger.py,sha256=1YALIaNYKTqeIyyCnYxzvZTK7x4FTSfYYl5CP9IMw8E,86
+gismap/utils/requests.py,sha256=DA-ifVMdcOtipDSqYdVRQi-7CGR5WCTYiGyZ7Xu78q0,1291
+gismap/utils/text.py,sha256=1_9DlduAYh7Nz-yAg-MaCTmdKbPPmuIY20bb87t7JAQ,3810
+gismap-0.2.0.dist-info/METADATA,sha256=XvYw05O-XgvC_5Ormcwjt4FOkbaJHHn-1pvwkV_ma8Q,2560
+gismap-0.2.0.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
+gismap-0.2.0.dist-info/licenses/AUTHORS.md,sha256=oDR4mptVUBMq0WKIpt19Km1Bdfz3cO2NAOVgwVfTO8g,131
+gismap-0.2.0.dist-info/RECORD,,

gismap-0.1.0.dist-info/RECORD

@@ -1,25 +0,0 @@
-gismap/__init__.py,sha256=TFBKB5AwAxzmcpIB48EpT4Ud9S417i5Obh85w2t1-Zc,598
-gismap/author.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-gismap/gismap.py,sha256=h0hwdogXGFqerm-5ZPeT-irPn91pCcQRjiHThXsRzEk,19
-gismap/gismo.py,sha256=kBzvlo-fqcvEgwQt-c_hHkfI1oTjBbKKb54vM8il_nI,6295
-gismap/search.py,sha256=nsUoDsFGeEtvCZ0dB7ooRPC_6qsazkiWx_oM7dHdNV4,4932
-gismap/lab/__init__.py,sha256=UQVY6Sk7o9O0YREArXA8_8ZfrDO0JYLy8tikSHBEbzo,377
-gismap/lab/graph.py,sha256=LOHNnmFqgsZc7YGUcFQ1GKl0DFiStFfoYqx-2IrjFAk,6525
-gismap/lab/lab.py,sha256=-mBcHcP9JaBAiiZlWRygac5MQdO0b6AriHWiLgTveWc,4224
-gismap/lab/lip6.py,sha256=yRx-RWgcGhGWUk_oyfgC9QYJSVnATnp6YlNvaOlr8Gw,1410
-gismap/lab/toulouse.py,sha256=KkzNOCFt3OU6pDJmDmli6wRNuMwH8CPXHj_S6aOcYeg,1462
-gismap/lab/vis.py,sha256=1QgLgLQjcB7rGiZ946lhLU1kR38oXp_Y8l8U-Ukn65k,5497
-gismap/sources/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-gismap/sources/dblp.py,sha256=5ZKngU66xgXffmiGHsbXGXJfU_NIybnuzXXZPcWIZoE,5381
-gismap/sources/hal.py,sha256=mWyiG12wXE9A8w81YGFicx8Yy9mX-anvIEoBueU2U6Q,9136
-gismap/sources/models.py,sha256=yJPBXcJO6MgOrSXgpGPepHDDtniJP1OsK8OSz__VzYc,543
-gismap/sources/multi.py,sha256=jfvBVgmnLT85RaEWcMNCZ8YtmSRuQuq8CkCHWY8u-hc,3515
-gismap/utils/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-gismap/utils/common.py,sha256=BV5mma4aikHyPIhH4theLkiNHaiqLv5_EfxhT0yajZo,1509
-gismap/utils/logger.py,sha256=1YALIaNYKTqeIyyCnYxzvZTK7x4FTSfYYl5CP9IMw8E,86
-gismap/utils/requests.py,sha256=DV5rmyuy7jpMp63kzGXe848pmwMb5397wGHgF0o_0BM,721
-gismap/utils/text.py,sha256=R7tKWevfF4g-LsohF6FeiRdCuNYz4jQ1VcQJIyhwei0,2393
-gismap-0.1.0.dist-info/METADATA,sha256=bTM9vZl81oJ5dUYakrCEL1uAbTZzEgC1AGSePg1TBU8,2499
-gismap-0.1.0.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
-gismap-0.1.0.dist-info/licenses/AUTHORS.md,sha256=oDR4mptVUBMq0WKIpt19Km1Bdfz3cO2NAOVgwVfTO8g,131
-gismap-0.1.0.dist-info/RECORD,,