Mexiko template for referense:

Template:
{{Photograph
 |photographer       =  <Personnamn / fotograf>
 |title              = 
 |description        = {{sv|<Beskrivning>. <Händelse / var närvarande vid>. <Motivord>}}
                                {{en|Translated <Händelse / var närvarande vid>}}
 |depicted people    = <Personnamn, avbildad>
 |depicted place     = {{city|<Places mapping>}} 
 |date               = <Fotodatum>
 |medium             =
 |dimensions         =
 |institution        = {{Institution:Statens museer för världskultur}}
 |department         = {{q:Q1371375|Etnografiska muséet}}
 |references         =
 |object history     =
 |exhibition history =
 |credit line        = 
 |inscriptions       =
 |notes              =
 |accession number   =
 |source             = The original image file was recieved from SMVK with the following filename:  <br /> 
'''<Fotonummer>.tif'''
{{SMVK cooperation project|COH}}
 |permission         = {{cc-zero}}
 |other_versions     =
}}
´´´
Categories:
[[Category:<Places mapping>]] AND/OR [[Category:<useful cat2>]]

OR
[[Category:Images_from_SMVK-EM_without_full_description]]
[[Category:Images_from_SMVK-EM_2016-11]]

Only five fotos in the metadata lacks description:

21 images in the metadata lacks photographer, see P4850

@Lokal_Profil Are these OK?

|photographer       =  {{Creator:John Lindros}} OR <empty> // this leaves out 6 images that have very little metadata anyhow. OK?

|description        = {{sv|<Beskrivning}} OR {{sv|<Nyckelord> // 6 cases with only the latter
                                {{en|The Swedish Cyprus expedition 1927-1931}}

@Mattias_Ostmar-WMSE

• Q: Is below correct? A: yes

|source             = The original image file was recieved from SMVK with the following filename:  <br /> 
'''<Fotonummer>.tif''' // Double check this!

@Lokal_Profil @Jopparn The images is currently licensed CC-BY-NC-ND (e.g. http://collections.smvk.se/carlotta-mhm/web/object/4258740) - should I ask Magnus to change this, like with the Mexiko images?

In T156612#2987416, @Mattias_Ostmar-WMSE wrote:

@Lokal_Profil Are these OK?

|photographer       =  {{Creator:John Lindros}} OR <empty> // this leaves out 6 images that have very little metadata anyhow. OK?

These should be fine. Only issue is how this affects the copyright if photographer is unknown. We'll check with MVK later today.

|description        = {{sv|<Beskrivning}} OR {{sv|<Nyckelord> // 6 cases with only the latter
                                {{en|The Swedish Cyprus expedition 1927-1931}}

English looks ok. For the swedish one I would suggest doing somehting similar to the Mexico collection. I.e.

|description        = {{sv| <Beskrivning>. <händelse>. <br /> ''Nyckelord:'' <nyckelord>}}

Where:

• <händelse> is the Swedish version of "The Swedish Cyprus expedition 1927-1931" (likely always the same string),
• "Nyckelord:" is only present if <nyckelord> is not empty.

As for whether to allow images without <Beskrivning> (but with <nyckelord>). If it's only 6 cases I would say upload them but tag with a maintenance category.

In T156612#2987435, @Mattias_Ostmar-WMSE wrote:

@Mattias_Ostmar-WMSE

|source             = The original image file was recieved from SMVK with the following filename:  <br /> 
'''<Fotonummer>.tif''' // Double check this!

With the addition of {{SMVK_cooperation_project|COH|museum=MM}}` underneath and correcting the typo (recieved ->received) I think that should work.

In T156612#2987442, @Mattias_Ostmar-WMSE wrote:

@Lokal_Profil @Jopparn The images is currently licensed CC-BY-NC-ND (e.g. http://collections.smvk.se/carlotta-mhm/web/object/4258740) - should I ask Magnus to change this, like with the Mexiko images?

Yes this needs to be updated before we upload.

In addition to the above:

• accession number should be {{SMVK-MM-link|1=foto|2=<Postnr.>|3=<Fotonummer>}}
• [[ Category:Media_from_the_National_Museums_of_World_Culture]] is likely not needed since it is added automatically by {{SMVK_cooperation_project}}
• department should be [[:d:|Medelhavsmuseet]]

• accession number should be {{SMVK-MM-link|1=foto|2=<Postnr.>|3=<Fotonummer>}} - No, SMVK-MM-link differs from SMVK-EM-link
• [[ Category:Media_from_the_National_Museums_of_World_Culture]] is likely not needed since it is added automatically by {{SMVK_cooperation_project}}
• department should be [[:d:|Medelhavsmuseet]]

In T156612#3016406, @Mattias_Ostmar-WMSE wrote:

• department should be [[:d:|Medelhavsmuseet]]

There wa a copy past missing in my original code. This should have been
[[:d:Q1331646|Medelhavsmuseet]] of course

In T156612#3016406, @Mattias_Ostmar-WMSE wrote:

• accession number should be {{SMVK-MM-link|1=foto|2=<Postnr.>|3=<Fotonummer>}} - No, SMVK-MM-link differs from SMVK-EM-link

Per an off-Phabricator discussion the correct format is {{SMVK-MM-link|1=<Postnr.>|2=<Fotonummer>}} (my bad)

Thanks, I noticed that.

I am discussing with SMVK about the licensing. They are looking into it, and are working on a broader policy change toward openness :-).

The infobox creation is done is the script create_infotexts.py, but the filename creation is done in metadata_to_json_and_fnamesmap.py.

Reopen since we need some further definitions.

|depicted place = {{city|<Places mapping>}}

I suggest we prioritize wikidata links here like so. Is this reasonable?

| depicted place ={{city|1=<wikidata item from places mapping|link=wikidata}} if exists OR ELSE {{city|<commons category from places mapping>|link=commons}}

However, the documentation had no examples of the use of the link parameter, so I hope this works as intended.

UPDATE: All places have wikidata items, so the result is that every image with <Ort, foto> matching the places will be mapped to wikidata and none to commons categories. Is this OK @Lokal_Profil?

Sometimes there are multiple people depicted in an image e.g. ´Lindros, John, Otterman, Gudrun, Westholm, Alfred´

The function flip_names() in batchupload.helpers() almost takes care of it, but would also need a split on every second ", " like e.g. so.

For reference; this is the pattern I used for |depicted people field since it was so few people to map:

def depicted_people_mapping(name_string_or_list):
    people_mapping = {
        "John Lindros":"[[Category:John Lindros|John Lindros]]", # [[:d:Q5957823|John Lindros]]
        "Lazaros Kristos":"Lazaros Kristos",
        "Alfred Westholm":"[[Category:Alfred Westholm|Alfred Westholm]]", # [[:d:Q6238028|Alfred Westholm]]
        "Erik Sjökvist":"[[Category:Erik Sjöqvist|Erik Sjöqvist]]", # [[:d:Q5388837|Erik Sjöqvist]] OBS! "Q" inte "K"
        "Erik Sjöqvist":"[[Category:Erik Sjöqvist|Erik Sjöqvist]]", # [[:d:Q5388837|Erik Sjöqvist]]
        "Einar Gjerstad":"[[Category:Einar Gjerstad|Einar Gjerstad]]", # [[:d:Q481299|Einar Gjerstad]]
        "Lazaros Giorkos":"Lazaros Giorkos",
        "Stefan Gjerstad":"Stefan Gjerstad",
        "Vivi Gjerstad":"Vivi Gjerstad",
        "Gudrun Otterman":"Gudrun Otterman",
        "Martin Gjerstad":"[[Category:Martin Gjerstad|Martin Gjerstad]]", # [[d:Q16632979|Martin Gjerstad]]
        "Knut Thyberg":"[[Category:Knut Thyberg|Knut Thyberg]]", # [[:d:Q16633505|Knut Thyberg]]
        "Rosa Lindros":"Rosa Lindros",
        "Ernst Kjellberg":"[[Category:Ernst Kjellberg|Ernst Kjellberg]]", # [[:d:Q5911946|Ernst Kjellberg
        "Bror Millberg":"Bror Millberg"
    }
    if isinstance(name_string_or_list, list):
        out_string = ""
        for name in name_string_or_list:
            out_string += people_mapping[name] + "/"
        return out_string.rstrip("/")
    else: # pre-supposes isinstance(name_string_or_list, basetring) == True
        return people_mapping[name_string_or_list]

if not item["Personnamn / avbildad"] == "":
    if len(item["Personnamn / avbildad"].split(", ")) <= 2:
        flipped_name = helpers.flip_name(item["Personnamn / avbildad"])
        mapped_name = depicted_people_mapping(flipped_name)
        infobox += "| depicted people    = " + mapped_name
    else:
        #print("Bökig | depicted person: {}".format(item["Personnamn / avbildad"]))
        words = item["Personnamn / avbildad"].split(", ")
        if len(words) % 2 == 0:
            span = 2
            list_of_names = [", ".join(words[i:i + span]) for i in range(0, len(words), span)]
            flipped_names_list = helpers.flip_names(list_of_names)
            #print(flipped_names_list)
            mapped_people = depicted_people_mapping(flipped_names_list)
            infobox += "| depicted people    = " + mapped_people
        else:
            print("Error: not even number of names in depicted people: {}".format(item["Personnamn / avbildad"]))
else:
    infobox += "| depicted people    = "
infobox += "\n"

Full code in create_infotexts.py

I decided to create commons categories for the people with wikipedia articles even though I don't know how many images there is for each one.

In T156612#3028489, @Mattias_Ostmar-WMSE wrote:

Reopen since we need some further definitions.

|depicted place = {{city|<Places mapping>}}

I suggest we prioritize wikidata links here like so. Is this reasonable?

| depicted place ={{city|1=<wikidata item from places mapping|link=wikidata}} if exists OR ELSE {{city|<commons category from places mapping>|link=commons}}

However, the documentation had no examples of the use of the link parameter, so I hope this works as intended.

I don't think you need the |link= parameter. Without it any [[ https://commons.wikimedia.org/wiki/Template:City | {{City|<Wikidata id>}} ]] should be resolved in the users language (linking to their Wikipedia). I don't think {{city|<commons category from places mapping>|link=commons}} will work since the allowed values are quite limited.

UPDATE: All places have wikidata items, so the result is that every image with <Ort, foto> matching the places will be mapped to wikidata and none to commons categories. Is this OK @Lokal_Profil?

From a {{City}} perceptive this is ok but do we still need the categories for categorisation of the images?

In T156612#3022074, @Mattias_Ostmar-WMSE wrote:

The infobox creation is done is the script create_infotexts.py, but the filename creation is done in metadata_to_json_and_fnamesmap.py.

I'll try to take a look at these in the coming days. Could you generate a few examples (stick them on a subpage on Commons?)

In T156612#3028593, @Mattias_Ostmar-WMSE wrote:

Sometimes there are multiple people depicted in an image e.g. ´Lindros, John, Otterman, Gudrun, Westholm, Alfred´

The function flip_names() in batchupload.helpers() almost takes care of it, but would also need a split on every second ", " like e.g. so.

Did you not write code to deal with this in the Mexico batch? Think I remember seeing it.

In T156612#3028814, @Mattias_Ostmar-WMSE wrote:

For reference; this is the pattern I used for |depicted people field since it was so few people to map:

    def depicted_people_mapping(name_string_or_list):
        people_mapping = {
            "John Lindros":"[[Category:John Lindros|John Lindros]]", # [[:d:Q5957823|John Lindros]]
            "Lazaros Kristos":"Lazaros Kristos",
            "Alfred Westholm":"[[Category:Alfred Westholm|Alfred Westholm]]", # [[:d:Q6238028|Alfred Westholm]]
            "Erik Sjökvist":"[[Category:Erik Sjöqvist|Erik Sjöqvist]]", # [[:d:Q5388837|Erik Sjöqvist]] OBS! "Q" inte "K"
            "Erik Sjöqvist":"[[Category:Erik Sjöqvist|Erik Sjöqvist]]", # [[:d:Q5388837|Erik Sjöqvist]]
...

I would stick these in a table on wiki rather than hard-code them. That way others can find the mapping in the future. If you want to avoid fetching from the wiki every time I would recommend loading a json file or similar.

else:
    print("Error: not even number of names in depicted people: {}".format(item["Personnamn / avbildad"]))

If this happens I believe it would be better to output the original name string and add the image to a maintenance category to clean it up manually afterwards then to throw it away.

@Lokal_Profil Thanks! Out of habit I added this while it's work in progress (wikitable coming up, printouts are for debugging now) and for documentation - this way it's easier to pick up the learnings after the batch is finished and some time has gone.

Row 456 in èxcel-export.xls` contains the row:

C04393+C04394 3924555 Svenska Cypernexpeditionen Teatern. Från öster. Sammansättningsbild. Soli. Svenska Cypernexpeditionen Cypern Soli 1927-1931 Lindros, John http://kulturarvsdata.se/SMVK-MM/Photograph/html/3924555`

This fotnummer, however is _not_ present in the google docs version (row is missing).

• check if a file named C04393+C04394.tifis present in the batch of images - and handle possible errors

wikitable people mapping is now added to commons

In T156612#3035926, @Mattias_Ostmar-WMSE wrote:

Row 456 in èxcel-export.xls` contains the row:

C04393+C04394 3924555 Svenska Cypernexpeditionen Teatern. Från öster. Sammansättningsbild. Soli. Svenska Cypernexpeditionen Cypern Soli 1927-1931 Lindros, John http://kulturarvsdata.se/SMVK-MM/Photograph/html/3924555`

This fotnummer, however is _not_ present in the google docs version (row is missing).

• check if a file named C04393+C04394.tifis present in the batch of images - and handle possible errors

I think C04393+C04394 is actually the photo number (at least according to http://collections.smvk.se/carlotta-mhm/web/object/3924555)

That said it's not impossible that the + has been replaced by another character in the filename.